The Application of Polymeric Reinforcement

in Soil Retaining Structures

NATO ASI Series
Advanced Science Institutes Series

A Series presenting the results of activities sponsored by the NA TO Science Committee,

which aims at the dissemination of advanced scientific and technological knowledge,
with a view to strengthening links between scientific communities.

The Series is published by an international board of publishers in conjunction with

the NATO Scientific Affai rs Division

A Life Sciences Plenum Publishing Corporation

B Physics London and New York

C Mathematical Kluwer Academic Publishers

and Physical Sciences Dordrecht, Boston and London
o Behavioural and Social Sciences
E Applied Sciences

F Computer and Systems Sciences Springer-Verlag

G Ecological Sciences Berlin, Heidelberg, New York, London,
H Cell Biology Paris and Tokyo

Series E: Applied Sciences - Vol. 147

The Application of Polymeric
Reinforcement in Soil
Retaining Structures
edited by

Peter M. Jarrett
Department of Civil Engineering,
Royal Military College of Canada,
Kingston, Ontario, Canada

and

Alan McGown
Department of Civil Engineering,
University of Strathclyde,
Glasgow, Scotland, U.K.

Kluwer Academic Publishers

Dordrecht / Boston / London

Published in cooperation with NATO Scientific Affairs Division

Proceedings of the NATO Advanced Research Workshop on
Application of Polymeric Reinforcement in Soil Retaining Structures
Kingston, Ontario, Canada
June 8-12, 1987

Library of Congress Cataloging in Publication Data

NATO Advanced Research ~orkshop on the "Application of Polymeric
Reinforcement in Soil Retaining Structures" (1987: Kingston, Ont.)
The application of polymeric reinforcement in soil retaining
structures / editors, Peter M. Jarrett, Alan McGown.
p. cm. -- (NATO ASI series. Series E, App11ed sCiences no.
147)
"Proceedings of the NATO Advanced Research ~orkshop on the
'Application of Polymeric Reinforr.ement in Soil Reta1ning
Structures,' Kingston, Ontario, Canada, June 8-12, 1987"--T.p.
verso.
"Published in cooperation .itM NATO Scientif1c Affairs Division."
Includes index.
ISBN-13: 978-94-010-7128-4
1. Geosynthetics--Congresses. 2. Soil stabilization--Congresses.
I. Jarrett. P. M. II. McGown, Alan, 1941- III. North Atlantic
Treaty Organization. Scientific Affairs Division. IV. Title.
V. Series.
TA455.G44N37 1987
624.1'64--dc19 88-14805
CIP

Published by Kluwer AGademic Publishers,

P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

Kluwer Academic Publishers incorporates the publishing programmes of

D. Reidel, Martinus Nijhoff, Dr W. Junk, and MTP Press.

Sold and distributed in the U.S.A. and Canada

by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.

In all other countries, sold and distributed

by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AH Dordrecht .. The Netherlands.

All Rights Reserved

ISBN-13: 978-94-010-7128-4 e-ISBN-13: 978-94-009-1405-6
DO.: 10.1007/978-94-009-1405-6

© 1988 by Kluwer Academic Publishers and copyrightholders as specified on appropriate

pages within.
No part of the material protected by this copyright notice may be reproduced or utilized
in any form or by any means, electronic or mechanical, including photocopying, recording
or by any information storage and retrieval system, without written permission from the
copyright owner.
 CONTENTS

Preface ix

INTRODUCTION

Reinforced Soil Retaining Structures and Polymeric Materials ......... 3

F. Schlosser and P. Delage

PREDICTION EXERCISE

Introduction and Rationale for Prediction Exercise 69

P.M. Jarrett

Laboratory Investigation of Two Large-Scale Geogrid Reinforced

Soil Walls........................................................... 71
R.J. Bathurst, W.F. Wawrychuk and P.M. Jarrett

Results of Class A Predictions for the RMC Reinforced Soil

Wall Trials .......................................................... 127
R.J. Bathurst and R.M. Koerner

An Assessment of the Resistance of TENSAR SR2 to Physical

Damage during the Construction and Testing of a Reinforced
Soil Wall ............................................................ 173
D.I. Bush and D.B.G. Swan

Preliminary Assessment of Sidewall Friction on Large Scale

Wall Models in the RMC Test Facility ................................. 181
R.J. Bathurst and D.J. Benjamin

Analysis and Predicted Behaviour for the Royal Military

College Trial Wall ..................•................................ 193
R.A. Jewell

CASE HISTORIES AND FULL SCALE TRIALS

Polymerically Reinforced Retaining Walls and Slopes in

North America ........................................................ 239
M.A. Yako and B.R. Christopher

Geotextile-Reinforced Retaining Structures: A Few

Instrumente,d Examples ................................................ 285
Ph. Delmas, J.C. Blivet and Y. Matichard

Geotextile Reinforced Retaining Walls: Discussion of

Instrumented Large Scale Test with Respect to the
Verification of Design Concepts ...................................... 313
B. Graf and J.A. Studer
vi

Review of Session on Case Histories and Full Scale Trials ............ 339

EVALUATION OF MATERIAL PROPERTIES

The Purpose of Materials Evaluation and Recommendations

for Materials Testing ................................................ 343
E. Leflaive and A. McGown

Review of Session on Evaluation of Material Properties ............... 357

Evaluation of Material Properties - Discussion ....................... 359

G.W.E. Milligan

ANALYTICAL TECHNIQUES AND DESIGN METHODS

Reinforced Soil Wall Analysis and Behaviour .......................... 365

R.A. Jewell

Reinforcement Extensibility in Reinforced Soil Wall

Design ............................................................... 409
R. Bonaparte and G.R. Schmertmann

Design of Reinforced Soil Retaining Walls: Analysis and

Comparison of Existing Design Methods and Proposal for
a New Approach ....................................................... 459
J.P. Gourc, A. Ratel, Ph. Gotteland

Factor of Safety Considerations in Reinforced Soil

Structures ........................................................... 507
R.T. Murray

Application of Finite Element Techniques to the Analysis

of Reinforced Soil Walls ............................................. 541
R.K. Rowe and S.K. Ho

Review of Session on Analytical Techniques and Design

Methods .............................................................. 555

Finite Element Analysis of Reinforced Soil ........................... 557

R. Chalaturnyk, D.H.K. Chan and J.D. Scott

Reinforcing Elements in Steep Slopes and Vertical-Faced

Earth Structures - German State of the Art ........................... 561
R. Floss

Analytical Techniques and Design Methods - Discussion ................ 569

G.W.E. Milligan

CONSTRUCTION METHODS AND ECONOMICS

Construction Methods, Economics and Specifications 573

C.J.F.P. Jones, A. McGown and D.J. Varney
 vii

Review of Session on Construction Methods and Economics .............. 613

RESEARCH NEEDS

Research Needs, Review of Session .................................... 619

K. Andrawes and P.M. Jarrett

PARTICIPANTS

Group Photograph 628

List of Participants ................................................. 629

Address List of Participants ......................................... 630

INDEX

Index 635
 PREFACE

Polymeric materials are being used in earthworks construction with ever

increasing frequency. The term "Geosynthetics" was recently coined to
encompass a diverse range of polymeric products designed for geotechnical
purposes. One such purpose is the tensile reinforcement of soil~. As ten-
sile reinforcement, polymers have been used in the form of textiles, grids,
linear strips and single filaments to reinforce earth structures such as
road embankments, steep slopes and vertically faced soil retaining walls.
A considerable number of retaining structures have been successfully con-
structed using the tensile reinforcing properties of "geosynthetics" as
their primary means of stabilization. Despite such successes sufficient
uncertainty exists concerning the performance of these new materials, their
manner of interaction with the soil and the new design methods needed, that
many authorities are still reticent concerning their use in permanent
works.

This book represents the proceedings of a NATO Advanced Research Workshop

on the "Application of Polymeric Reinforcement in Soil Retaining Struc-
tures" held at the Royal Military College of Canada in Kingston, Ontario
from June 8 to June 12, 1987. The initial concept for the workshop occur-
red during the ISSMFE Conference in San Francisco in 1985 when a group of
geotextile researchers mooted the idea of hoiding a "prediction exercise"
to test analytical and design methods for such structures. It was decided
that members of the group and others would attempt to predict in advance
the behaviour of two, 3m high, polymerically reinforced soil walls to be
constructed and tested in the laboratories at the Royal Military College.
After the tests were completed the group would meet to discuss and compare
predictions and behaviour in order to see where the results pointed
concerning our analytical abilities. The "prediction exercise" began. In
seeking the means to get this international group together it was recog-
nized that a NATO Advanced Research Workshop might be appropriate. Early
in our approach to the NATO Scientific Affairs Division it was realized
that their sponsorship would allow us to broaden the scope of and atten-
dance at the workshop. Thus the "prediction exercise" became the initial
focus and springboard for a group of experts to study and thoroughly
discuss the state of the art of this new form of construction. This volume
contains the results of the prediction exercise, the papers developed by
the leaders of five sessions in which all aspects of this method of con-
struction , were addressed and summaries of certain of the session
discussions including the research needs session.

The workshop ultimately drew together 37 researchers, builders, designers

and manufacturers from 12 countries. The workshop fell under the Double
Jump program and was attended by six persons representing manufacturers of
geosynthetic products. The harmon ius manner in which this diverse group
ix
x

worked together was quite outstanding and went a long way to ensuring that
the scientific objectives of the workshop were met. The extension of the
working cooperation to the social aspects of life was equally outstanding
leading for example to an impromptu international "soccer" match amongst
participants that showed up some unexpected yet quite prodigious skills.
With such spirit and congeniality it is felt that the additional objectives
of the NATO Scientific Affairs Division for the establishment of inter-
national understanding and cooperation amongst scientists were also ful-
filled at this workshop.

To reach the position where the scientific group could meet in such produc-
tive circumstances and have a book produced from the proceedings demands
considerable effort from a number of people and financial and material
support. The Scientific Director for the meeting and the editors of this
volume wish to acknowledge with sincere thanks the following:

The NATO Scientific Affairs Division for their financial support of

the Advanced Research Workshops. This far-sighted aspect of NATO is
now far better appreciated by the group of scientists involved in
this meeting.

The Department of National Defence of Canada who through the Royal

Military College of Canada provided the meeting facilities and con-
siderable material and administrative support to enable the workshop
to be 'organized.

The Workshop Organizing Committee of Drs. Andrawes, Leflaive and

Koerner who leant weight and enthusiasm to the original propoual and
wise counsel throughout.

The session leaders and authors for their co-operation in the

presentation of papers and submitting them by the deadlines or at
least reasonably close to deadlines in some cases.

Mr. J.E. DiPietrantonio Qf RMC for his considerable work and skill
in preparing, repal.rl.ng and mounting of drawings in many of the
final edited papers.

Miss Martina M. Lahaie of RMC for her considerable secretarial and

administrative skills and her long term enthusiasm to see the work-
shop succeed. Her contributions prior to and during the workshop
and in the editing of the manuscripts for this volume were the
keystones of the administrative process.

In conclusion, the workshop occurred through the administrative efforts

described above but it suceeded scientifically through the cooperative
efforts of all the participants. They were a pleasure to do business with!

Peter M. Jarrett Alan McGown

Royal Military College University of Strathclyde
Canada Scotland
Introduction
REINFORCED SOIL RETAINING STRUCTURES AND POLYMERIC MATERIALS

F. SCHLOSSER and P. DELAGE

Ecole Nationale des Ponts et Chaussees, CERMES

B.P. 105 - 93194 NOISY-LE-GRAND CEDEX

1. INTRODUCTION
Soil has been used since the most ancient civilizations as a
construction material. However, due to insufficient tensile strength,
designers had to improve its resistance by using mechanical processes
(compaction, draining), chemical processes (stabilization), or by in-
clusion of resisting elements (reinforcement).
In the past, natural inclusions have most often been used to improve
the mechanical properties of soils and structures straw in clayey
soils to develop construction material; palms or branches at the base
of structures founded on soft soils and so on. It is also interesting to
mention the reinforcement of natural slopes provided by plant roots in
the nature.
During the past twenty years, the important development of Reinfor-
ced Earth, and the concept of reinforced soil as a construction mate-
rial, introduced by its inventor H. Vidal, in the sixties, have
contributed to the birth of a new area of soil improvement soil
reinforcement. This area is based on a generalization of the "reinforced
soil" concept, including various techniques, {or slopes and embankments,
retaining walls, foundations, dams .••
The concept of soil reinforcement is based on the existence of a
strong interaction between the soil and the inclusion. The most common
interaction is friction, but passive pressure may also be mobilized.
Frictional interaction requires good mechanical properties of t~e soil,
particularly in terms of friction angle and drainage; granular soils
are best adapted for this kind of reinforcement. Passive pressure is
generally associated with anchors, which needs more sophisticated in-
clusions. It can be mobilized in all types of soils, including saturated
fine-grained soils, where frictional properties are poor.
The number, type, and arrangement of the inclusions may be quite
variable. Depending on the type of the inclusion, two extreme cases may
be considered (Schlosser et aI, 1983).
1) a "uniform inclusion", for which the soil-reinforcement interac-
tion can develop at any point along the inclusion. In this case, a
relatively high and uniform density of reinforcements will result in a
new composite material called "Reinforced Soil". The behaviour of the
reinforced soil mass can be investigated through laboratory testing
considering a representative sample of the new composite material. This
concept is illustrated in figure 1.

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 3"'{)5.
© 1988 by Kluwer Academic Publishers.
4

• Composite
Uniform
__- - - - - - - - 1 inclusion
inclusion
I---------------~ Anchor

1---------./
Reinforced Earth wall Ladder wall
I) - REINFORCED SOIL 2L MULTI ANCHORAGE SYSTEM
@,.. PERIODICAL REINFORCEMENT

.. ./ . · . . i
Granular
material Geotextile
--,

~.. ' '. L' .. :.'. , ~

s.z;:::i:;0it)'S:::::.:: ~.;~ i.s~:.'-

..------- - - - - -
_ - :... _. • Soft soil -- -

I) - MEMBRANE 2)- PILE

© - ISOLATED REINFORCEMENT

FIGURE 1. Types of soil reinforcement systems (Schlosser et al., 1983)

2) a "composite inclusion", which consists of an inclusion reinforced

at particular locations where the soil-reinforcement interaction is
concentrated. Generally, as for multi-anchorage systems, these points
are located at the ends of the inclusions.
These considerations lead to the classification of soil reinforcement
systems presented in Table 1.

~
forcement uniform composite
Density
reinforcement
multi _
periodical re i nforced anchorages
soi I systems

isolated membranes anchorages

piles

Table I - Classification of soil reinforcement systems

At present time, the development of polymer reinforcements, which are

essentially two dimensionnal (geotextiles, geogrids, geomembranes) could
 5

lead to a modification of this classification, and to a distinction

between linear and two dimensionnal reinforcement.
For Reinforced Earth strips, various materials were initially used
(Schlosser, 1977) : fiberglass in a polyester resin, Tergal, or passiva-
ble metals such as aluminum or stainless steel. Finally, metal, and more
particularly galvanized steel, was elected because of low deformability,
cost and ease of setting.
In the early seventies, the use of polymers started to develop,
mainly through geotextiles. Progressively, designers were attracted by
the great variety of available products. They also considered polymers
as an alternative, with respect to metal corrosion problems.
Numerous papers have been published on polymer reinforcement in
different geotechnical journals and conference proceedings.
Among the conferences partially or totally devoted to soil improve-
ment, those related to the use of polymers in reinforced soil retaining
structures are listed below :
1) Soils and fabrics, Paris, 1977.
2) Reinforced Earth and other composite soil techniques, Edinburgh
1977.
3) Earth Reinforcement, ASCE, Pittsburgh 1978.
4) International Conference on Soil Reinforcement, Paris, 1977.
5) 2nd International Conference on Geotextiles, Las Vegas, 1982.
6) 8th ECSMFE, Helsinki, 1983. General Report on Soil Reinforcement.
7) Symposium on Polymer Grid Reinforcement in Civil Engineering,
London, 1984.
8) 11th ICSMFE, San Francisco, 1985 : General Report on Geotechnical
Construction.
9) 3rd International Conference on Geotextiles, Vienna 1986.
A journal exclusively dealing with the subject, entitled "Geotextiles
and Geomembranes", is now regularly published.
This paper will review the development of reinforced soils systems
for retaining walls, with a special attention to the use of polymeric
materials such as geotextiles, geogrids and fibers. The history and
development of soil reinforcement systems for retaining walls are first
presented, after which the main types of existing polymer reinforced
soil walls are described. The properties of polymers in relation to the
behaviour and the design of reinforced soil walls are briefly consi-
dered. Finally, advantages and drawbacks of using polymers in s~ch walls
are discussed.

2. HISTORY AND DEVELOPMENTS

2.1 Coyne's ladder wall
In 1929, Andre Coyne patented in Paris a multi-anchorage system to be
used for the construction of retaining walls and especially quaywalls,
dykes and so on. The idea of using such a system was the principle of
constructing a wall by successive horizontal elements, formed by a light
facing element linked to continuous or discrete anchors with ties. In
this system, the ratio between the total height of the wall, and the
length of the anchor ties was approximately 2.5. Figure 2a shows a
schematic view of such a structure as specified in the patent, and
Figure 2b a cross section of an actual quaywall, 200 m long, built in
Brest harbour in 1928, which was the first major application of the
system. Despite some studies performed on reduced scale models (Coyne,
1945), the real mechanism of the ladder wall was not fully explained by
Coyne. He indicated that the structure formed by the facing, the tie-
6

rods, the back-fill material located between this facing and the ancho-
rages may apparently behave as a solid, sustaining small deformations,
as compared to the displacements of the wall. Coyne (1945) describes
some applications of this patent: the work presented in figure 2b is a
dyke submitted to tides, which may even be submerged by big waves
internal seepage may then occur, with 1,50 or 2 m hydraulic gradients,
and the structure behaves as a dam; good rocky material was used as a
fill material between the ties ; this structure supported 0,50 m settle-
ments without any problems, because of good articulations between the
facings. Coyne also mentions a 200 m long quay wall in Brest, and the
construction of 10 to 20 m high retaining structures, such as the side-
walls of an overflow in Mareges, a 10 m high coffer-dam at the same
location, and a 14 m high dam on the river Laurenti (Pyrenees) ; in the
case of dams, the ladder wall may constitute either the up-stream or the
down-stream facing. In all cases, good rocky material is placed in
contact with the ties.

+8.00 Anchor beam

E
o
10
..,.
a Ordinary
backfi II material

® Patent figure ® Ladder wall at Brest (1935)

FIGURE 2. Ladder wall system invented by Coyne (1929)

Apparently, the first developments of this technique were stopped

during World War II, and, in spite of Coyne's wishes, it was not suffi-
ciently used later during the reconstruction of the country.
After a long period of time during which the Ladder Wall system was
not used, several systems related to this type of reinforcement have
been proposed since the starting point of Reinforced Earth development.
They differ from each other with respect to the type of anchorage and
facing used. Anchored Earth (Murray and Irwin, 1981) and Microanchorages
(Costa Nunes, 1978) use frictional anchorages, while American Geotech
and other systems use passive pressure beams or plate anchors (Figure
3) •
 7

Anchored earth (Murray, 1981 ) American geotech

Micro-anchorages (Costa Nuiies,1978)

FIGURE 3. Reinforcements in recently developed ladder type walls

It is interesting to mention a mixed system developed by Fukuoka et al.

(1982), in which the facing used is made of fabric att a ched to vertical
columns and the anchors are concrete plates. The type of behavioral
mechanism involved in this system was demonstrated with a full-scale
e x periment: the displacement (rotation) of the rigid columns is suffi-
cient to reach the value of the active earth pressure on the facing,
whereas the pressures on the anchor plates remain equal to the Ko state
of stress.
Chabar et al. (1983) described the construction of a 21 m high dam,
built according to the classical Coyne's ladder wall system.
More recently a multi-anchorage system, called Actimur (1984) has
been proposed in France. It combines a vertical sheet-pile tacing and
horizontal tie-rods with vertical metallic anchor discs.

2.2 Reinforced Earth

The inventi~f Reinforced Earth by Henri Vidal in 1963 and the
rapid development of this new technique at the end of the 60's, has been
the starting point of reinforcement systems, especially dealing with
soil retaining techniques where the soil reinforcement is periodical and
where the soil-reinforcement interaction acts all along the reinforce-
ment. These systems have been denominated reinforced soils by Schlosser
et al. (1983).
Henri Vidal considered Reinforced Earth as a new composite material
and consequent.l y introduced the very interesting concept of reinforced
soil material, which has prooved to be general, realistic and efficient.
8

@ Combination of reinforcements alone

@ Concrete

© Macro molecules and human body materials

FIGURE 4. Cohesive materials as combinations of "grains" and reinforce-

ments CR. Vidal, 1966)
 9

It must be noticed that, in his first paper (1966), H. Vidal developed a

large theory, presenting the different manners of producing a cohesive
material using independant grains and reinforcements. As indicated in
figure 4 he dealt first with the texture of reinforcements made with
fibers (nonwoven, wooven, etc ••. ) and explained the behaviour of several
materials (wood, paper, clay, concrete and finally human body materials)
by associations of "grains" and reinforcements interacting through fric-
tional forces.
An experimental study of the behaviour of the Reinforced Earth mate-
rial was performed at the Laboratoire Central des Ponts et Chaussees
(Schlosser and Long, 1972) by testing samples of sand reinforced with
horizontal and regularly spaced aluminium foil discs in the triaxial
apparatus. It was shown that two failure modes can develop in such
reinforced sand samples: failure by slippage of the reinforcement, and
failure by reinforcement breakage. The yield line in the (a l , a 3 ) prin-
cipal stresses axis is presented in figure 5: at low conflning pres-
sures, failure occurs by slippage, leading to a curved yield line pas-
sing through the origin; at higher confining pressure, this failure
line is a straight line which proves that the reinforced sand behaves as
a cohesive material having the same friction angle as the original
sand and an anisotropic pseudo-cohesion due to the reinforcements. This
pseudo-cohesion is very rapidly mobilized at low axial deformations,
since the reinforcements behave in a rigid way compared to the relative-
ly deformable sand. Tensile stress measurements using strain gauges show
that the maximum stress value in the discs is obtained at points located
at a distance from the center approximately equal to two-third of the
radius. The inclined failure plane developing when the sample fails by
"reinforcement breakage", indicates that a bifurcation phenomenon occurs
in the development of the failure surfaces.
A very interesting theoretical contribution to the subject is due to
Bassett and Last (1978). These authors considered that the mechanism of
tensile reinforcement involves anisotropic restraint of the soil defor-
mations in the direction of the reinforcements. Then they used Roscoe's
failure criteria for sands, based on zero extension concept, to demon-
strate that the presence of the reinforcements leads to a rotation of
the principal directions of the deformations tensor. They showed (figure
6) that since in a Reinforced Earth wall, the direction of the reinfor-
cement must be aligned with the zero extension direction, the. failure
surface must be vertical to comply with the assumption of suppressed
dilation rate (it is assumed that the Reinforced Earth material exhibits
a zero dilation angle). In other words, it can be said that, due to the
soil-reinforcement interaction, the presence of the reinforcements in a
soil mass greatly modify the strain and stress patterns. Moreover this
is consistent with the development of cracks along a cylindrical surface
in reinforced sand samples in the triaxial apparatus.
10

BREAKAGE
.
30
.
20 --- E

-
u
0"3:::: 2
2cml
j
10
(Aluminium
.. foil

2 3 4 5 6

@ Reinforced Earth material fai lure I ine (Schlosser and Long,1972)

CTI

5300

High CT,

Low 0'2

CD Active zone
® Resis!ant zone r/R
- 700 L-JL..-.L-_ _L-----,-.L-__':-::------,L---:-.L
L--------r--------~ 1 9/0 2/3 1/3 k'O 1/3 2/3 9/01

(FEM calculations) (Experimental)

@ Tensile stresses distribution and crack locations in the discs

FIGURE 5. Behaviour of the reinforced earth material at the triaxial ap-

paratus
 11

a Li ne -!;;Ft+-t=j.t-t~r+r.r;,::-~_- Pot ent ial

failure line
f3 Line

FIGURE 6. Influence of the reinforcement on the potential failure lines

(Basset and Last, 1978)

Based on the above principles, Reinforced Earth consists essentially

of the following components : 1) a granular backfill material, 2) linear
reinforcements, generally strips 3) a facing made of pre-cast elements
attached to the strips. The two major components are the granular
backfill and the strips; the purpose of the facing is only to retain
locally the backfill between two horizontal reinforcement layers. Figure
7 shows a schematic view of a Reinforced Earth wall, showing these three
components.

Backfill material of
good quality
strip_

FIGURE 7. Schematic view of a reinforced earth wall

12

There has been a great improvement in the technological development

of the Reinforced Earth technique, with respect to all three components,
i.e. the facing, the strips and the backfill material. Initially the
facing was made of U-shaped elements, 33 cm high, the weight of each
being light enough to enable an easy handling. The strips were complete-
ly smooth, generally 60 mm wide and 3 mm thick. The backfill was a good
granular material with less than 15 % in weight of grains smaller than
80 ~m (n° 200 sieve). Considering the purpose of the present paper, it
is interesting to note that H. Vidal first planned to use plastic strips
and plastic facing elements in order to avoid corrosion problems, as
indicated in his first paper (1966). He was rapidly able to produce in-
dustrially in 1967 facing elements and strips made of fiber-glass coated
with polyester resin. It will be seen further why the use of such a
material was stopped and replaced by metals.
Three events have marked the Reinforced Earth technological develop-
ment. Firstly, the choice of galvanized steel for strips and facing,
after a first tentative with polyester coated fiber-glass and stainless
steel and aluminum used for some years in France. These two metals were
at that time (and even now) considered to be particularly efficient
against corrosion, even when embedded in soils. They are theoretically
protected by a thin layer of undestructible oxyde on their surface.
However, it must be now accepted 15 years later that this is absolutely
not true and that these metals may be corroded in some cases more
drastically and rapidly than galvanized steel, due to a special phenome-
non involving an accelerated corrosion rate.
The second event has been the development in 1971 of a typical cruci-
form panel for the facing. This type of facing enables architectural
possibilities, curved facings and it is now worldwide representative of
the Reinforced Earth development.
In 1975, the Reinforced Earth Company patented the ribbed strip. This
new technological aspect was directly issued from research on the soil-
reinforcement frictional interaction. As indicated later by Schlosser et
Elias (1978), the main phenomenon in this 3-dimensional friction
mechanism is the restrained dilatancy effect (figure 8). The
consequently apparent friction coefficient is much influenced by the
volume of the sheared soil zone around the strip. Ribs increase this
volume and thus largely increase the ~* value.
After 20 years, the Reinforced Earth major developmen~ appears to be
related to following features: 1) R.E. behaves satisfactorily even in
various critical situations (large differential settlements, movements
in the foundation soil, seismic event, etc ..• ). 2) R.E. cost is competi-
tive and generally low compared to other solutions. 3) R.E. wall facings
are attractive and aesthetic.
For the time being, the only problem is related to the special corro-
sion of the stainless steel and a luminum strips embedded in walls built
in France 10 to 15 years ago.

2.3 Geotextiles
The use of geotextiles in earthworks for reinforcement and separation
at ,the base of an embankment on soft soil started approximately at the
same period as the Reinforced Earth early development. In fact, the
first paper dealing with such an application has been published in 1969
(Vautrain and Puig).
 13

JL"
7

_ 6 \
Ribbed reinforcement
.,c
] 5
~
o GRAVEL
04 • T
Y=21 kN/m 3
c
o JL = n.
~ <p =46° (soil/soil)
...
';: 'i' =27,5° (soil/reinforcemen

~2
&.
0.1
_
---"4-
Smooth reinforcement
----e__ tg<p
<t
_ - - -- - - - - - - --:::::!:=-=.--:.::=:-----___ - - tg'i'
O'L----~----2~----~3r-----~4c-----~----~6~---
Height of embankment above the reinforcement - H (m)

® Restrained dilatancy effect mechanism

(kPa)
. "
.....

100
CT v 1>- PRESSURE CELLS

dr= ~-ANCHOR
50 20cm

ero
Distance d (em)
2816 70 100 200

® Normal pressures measurements around a tensioned inclusion

( Plumelle, 1984)

FIGURE 8. Restrained dilatancy effect on soil-linear inclusion friction

14

Since this time the application of geotextiles to roadways,

embankments and slopes has intensively increased. According to Giroud
and Carroll (1983), the largest quantity of geotextiles is now utilized
for roadway construction, principally temporary and construction roads.
The first application of geotextiles to multilayered soil-fabric
retaining systems was done in 1971 (Puig et aI, 1977). It was an experi-
mental wall using a nonwovp.n fabric (Bidim) and a very poor backfill
material (wet clayey and sensitive soil). The wall was 4 m high and was
founded on a very compressible soil (peat layer, 3 m thick).
Since this first application, geotextiles have been used for
retaining walls, and for earth dams (Kern, 1977). They present interes-
ting features: low cost, drainage, possibility of using poor backfill
material. However their utilization has been rather limited until now,
probably because of their deformability (particularly in the case of
unwoven geotextiles) and to the relatively inaesthetic appearance of the
facing.

2.4 Grids
T~first reinforced soil retaining structure uSing grids as
reinforcements was constructed in 1974 on Interstate Route 5, near
Dunsmuir, California (FORSYTH, 1978). One year before in 1973, the
California Transportation Laboratory developed a large direct shear
device in order to test the pullout resistance of different
reinforcement systems (smooth strips, ribbed strips, bars, bar mats).
The purpose was to find a reinforcement system which could enable the
use of granular backfill material containing a large percentage of fine-
grained material. It was found that the best system-was the grid or bar
mat, which provides a relatively linear reinforcement, wisthstanding
large pull-out forces, thanks to the passive thrust mobilized against
the transversal bars. However, compared with the pure frictional
interaction reinforcements, i.e. strips, these bar mats require large
displacements, 5 cm and more, to fully mobilize the pull-out resistance
(CHANG et al., 1977).
Figure 9a shows the welded bar mat tested and used by Cal trans in the
construction of the first "mechanically stabilized embankment" at
Dunsmuir. This structure was approximately 120 m long and consisted of
two walls, 6 m high. The vertical facings were made from rectangular and
long precast concrete elements 3.75 m X 0.6 m X 0.2 m. in size (figure
8c) . The bar mats, which were 1.2 m wide and 3 to 4.5 m in length, were
attached to the facing elements by inversion of the two-bar yoke through
precast holes in the facing panels. These two prethreaded bars were
bolted into position, inducing some interesting prestressing of the
reinforcement. The construction was very similar to Reinforced Earth.
As indicated by Forsyth (1978), it was "anticipated that the bar- mat
mode of reinforcement would have significant economic advantage in
certain areas of the state of California where high quality backfill
material is not readily available".
This argument was considered and put forward by VSL when promoting in
1980 an equivalent retaining system, called Retained Earth, in which the
same , type of welded bar-mat was used.
As shown by Schlosser et al. (1983, 1985), the bar-mat interaction
mechanism is complex and involves both friction along the longitudinal
bars and passive thrust again the transversal bars. For small soil-
reinforcement displacements « 0.5 cm) there is initially a mobiliza-
tion of the friction along the longitudinal bars. For larger displace-
 15

.1 I.Gem

@ Welded bar mat @ Special shear device

(California Transportation Laboratory)

© Prefabrication panel used in Dunsmuir wall

FIGURE 9. Type of welded bar mat and panel used by California trans-
portation (Cal trans)
16

ments there is a mobilization of the passive pressure on the transverse

bars and the stress-displacement curve keeps increasing even for displa-
cements greater than 10 cm.

Bars (L= 8m ; 0= 0,o3m) Discs

cr T !._ ..... -._ ... ( D=0,3m)
(kPa) (kPa)
2000 100

FRICTION

---+-
Sand", =40°
Depth =3m
1000
~ Two discs bars
PRESSURE + One disc bars

~+
400
+

200
Head displacement
5 10 15 20 25 8(mm)

@ Respective mobilization of friction and passive pressure on bars

equipped with anchoring discs (after Morbois and Long, 1984 )

F (kN) Pull out force

100
Complete mobilization

( " h;"""

I GRAVE L I
Displacement 8 (cm)
o'L-------~5------~1~0------~15~----~2~0~----~2~5·

@ Pull out test on a grid (Bacot,19SI)

FIGURE 10. Mechanism of bar pull-out resistance

 17

Figure 10 shows two typical results about this phenomenon 1) the

respective mobilization of friction and passive pressure on bars equip-
ped with anchoring discs ; 2) the pull-out force-displacement curve on a
grid.

@ PERFORATED
SHEET

... Roll length

@ UNIAXIAL
GEOGRID
(5
0::

- -
....
- -
Roll width

Junction
£:
® BIAXIAL
GEOGRlp
~
c:

(5 Rib
0::

I.. 3.80cm .1

FIGURE 11. Typical geometry of perforated sheet and uniaxial and biaxial
geogrids (Bonaparte et al., 1984)
18

Because of this mechanism, bar-mats are more resistant in pull-out

than frictional reinforcements (bars, strips) only for large displace-
ments (5 to 10 cm). If such lateral displacements values are allowable
for the structure, it appears that bars-mat reinforcements permit to use
poor quality backfill material with a large percentage of fine-grained
soil in a retaining system. However, further research in this field is
still required to specify suitable soils and what may be the best mat
geometry.
In the early 80's, Netlon manufactured and developed a plastic grid
product called Tensar. This material consists of a high strength, orien-
ted polymer grid structure obtained from punched and stretched polymer
sheets. The rapid development of this product, used in a variety of soil
reinforcement applications (embankment reinforcement, retaining walls,
rafts, repairs of slope failure, gabions), led to a new type of
tridimensional reinforcements called geogrids. A geogrid has a small
opening size (about a few centimeters) compared with a bar-mat (10 cm
and more), but compared with non woven geotextiles, it exhibits a larger
deformation modulus and tensile resistance (figure 11).
Properties and applications of geogrids will be further discussed.
With respect to soil-grid friction, there is some similarity with the
soil-bar mat interaction. However, the mechanism involves a new phenome-
non in coarse granular soils, resulting from interlocked soil particles
within the grid apertures, which act as an anchor for the transverse
ribs of the grid. Forsyth and Bieber (1984) have compared a plastic grid
(Tensar) (mesh size of a few centimeters) with a metallic bar-mat (20 cm
opening size), both having identical surface areas. The force-displace-
ment curves obtained in pull-out tests for a normal stress equal to 34.5
kPa are proportional with a ratio of about 3, because the passive pres-
sure effect observed on transverse elements is lower for the Tensar
(Figure 12). The type of soil used was decomposed granite (~' ~ 35°).

DECOMPOSED GRANITE: <P ~35°

o;=34,5kPa

_100
z
-'" Metallic bar mat.
LLI failed by slippage
u
cr 75
0
lJ..
I-
:::>
0
50
..J
..J
:::>
a.
Netlon ER 2 (Tensor)
25 Failed by
tensile break

0.5 1.0 1.5 2.0 2.5 3.0 3.5

6L DISPLACEMENT (em)

FIGURE 12. Pull-out force/displacement curves for a metallic bar mat and
a plastic grid
 19

However, comparing rapid direct shear tests on clay samples transver-

sally reinforced by a metallic or plastic grid, Jewell and Jones (1981),
and Ingold (1983) found no difference between plastic and metallic
grids. The reinforcement effect on the undrained shear resistance was
the following: 12 % increase with smooth steel sheet, 37 % increase with
corrugated steel sheet, 42 % increase with steel grid and 44 % increase
with the geogrid.
It must be noted that there is a great difference between the two
types of tests: pull-out test and Jewell's direct shear on transversal-
ly reinforced samples. The latest is more representative of the friction
phenomenum close to the potential failure surface, for instance in a
reinforced soil retaining wall. However when using long reinforcements
beyond this surface, as generally recommended for a design, the first
type of test appears more adequate, according to the authors opinion
(figure 13).
According to Bonaparte et al. (1984) high density polyethylene· or
polypropylene are suitable for soil reinforcement because of their in-
ground durability and resistance to chemical, as well as micro-
organisms attack. Generally speaking, durability is one of the most
important problems, because reinforced soil structures are alternatives
to classical reinforced concrete structures and they must therefore
present an equivalent service life. Durability of polymeric materials
will be treated in a subsequent section.
At present time, geogrids have largely been used in embankment
reinforcements, rafts, gabions and corrections of landslides, but only a
small amount of retaining walls have been built. It seems that like for
geotextiles, the problem related to the facing still need to be solved:
inaesthetic aspect and erection difficulties in geogrid facings,geogrid
attachment to the panels in prefabricated facings.

2.5 Three-dimensional reinforcement : Texsol and fibers

- ------ --- ------
It is technically very difficult to obtain in-situ soil reinforcement
with flexible inclusions oriented in all directions as it is the case
with the natural in-situ reinforcement provided by roots. However, the
development of geotextiles and synthetic fibers has raised interest for
three-dimensional reinforcement of soil fills, a process in which the
soil is mixed with small inclusions (fibers, small plates) or continuous
filaments (Texsol). The materials obtained in this manner, in which many
particles are bound to each-other by the reinforcements, have a
structure resembling concrete, rather than one or two dimensional mate-
rials. Their behaviour mechanism is complex and has not been
sufficiently studied until now. Despite their interesting potential use,
applications are still limited.
The principle of the patented Texsol consists in placing within the
mass of a granular soil one or several continuous filaments (Leflaive
et al., 1983), in order to obtain a three-dimensional random mixture of
filaments and soil particles. The inclusion is a polyester filament,
with a diameter 0.1 mm and a tensile strength of about 10 N. The
placement technique involves a simultaneous projection of sand, water
and filament, as indicated in figure 14. In spite of a low percentage by
weight of the filament (0.1 to 0.2 %), the total length is very impor-
tant (about 20 cm per cm 3 of material), which explains the good mecha-
nical behaviour at failure.
20

Potential failure
i-- surface

® ~ CO ~~
~ ~
r
\
\
\
\,
'\
"'-

JEWELL'S 01 REeT
SHEAR TEST

.,: ",",.' .'

® . ',; .. '.:........ :.:-. .. '

PULL- OUT
:
,' ,",",",

.:-.. : :' .~ ~.~ ::::: ',:' :',:.:.,

(. TEST
. :. :. ': ' . ~' .. ': : ..; : .'.
\

FIGURE 13. Differences between Jewell's direct shear test and pullout
test
 21

Figure 14 shows the first job carried out with this material
(Leflaive et al., 1983). The problem was to repair a slide in a cliff of
chalk and to restore a path at the top of the cliff. About 300 n~ of
Texsol were placed, at a slope angle of 60° and with difficult access
conditions.

"T".....,~-t-l/ Existing wall

Path ta be restored

1
Texsol embankment
Thread reel
6m

CHALK
Placement of the material
6m

FIGURE 14. Texsol embankment of Caudebec-en-Caux and placement of the

material (Leflaive, 1983)

At present time, Texsol is essentially used for retaining walls, but

some very interesting applications could be found in relation with
foundations, slabs or road construction.
The presence of continuous filaments increases the resistance and the
ductility of the sand. A real cohesion of about 100 to 250 kPa is
2reated. However, due to the placement by approximately horizontal
layers, this cohesion could be relatively anisotropic with a small value
on horizontal planes.
In fact, only very few papers have been published on the mechanical
behaviour of Texsol. CBR tests, unconfined compression tests, static and
dynamic triaxial tests have been carried out (Leflaive et al., 1983),
but no stress-strain curves and no details about the preparation of the
samples have been given. It is interesting to notice that the relative
density of the sand has a main effect on the cohesion value for a
loose sand this value is smaller than 100 kPa, for a dense sand, it
ranges from 150 to 250 kPa. But the way this cohesion is mobilized with
the axial deformation in a triaxial test, as it has been analysed for
sand samples reinforced by discs, remains a fundamental question. As
reinforcement by filaments do not increase the deformation modulus, it
could be expected that this mobilization would require relatively large
deformations.
The inventor of this technique, Leflaive, says that besides the
friction interaction between the soil and the filaments, two other
phenomena are involved in the reinforcement mechanism with a continuous
filament :
22

1) The curvature effect of the filament, associated with the fric-

tion, considerably increases the resistance to sliding.
2) The three-dimensional entanglement effect, coating each volume of
soil with a large number of threads.
The curvature effect is partly similar to the inclusive undulation
effect studied by Bacot (1981). This author has experimentally and
theoretically studied the influence of the undulation when using linear
and flexible inclusions. Such an undulation is mostly characteristic of
reinforcement of fills (reinforced earth, embankment reinforcement, mat
foundations, etc ••. ) for which the inclusions are set in place layer
after layer on an embankment surface which is not perfectly plane. The
variation of the apparent friction coefficient ~* as a function of the
total angular variation Bt of the inclusion with respect to the.horizon-
tal, is exponential. Bacot considers that this effect could be added to
the dilatancy effect, which would explain the observation of high values
of ~*, higher than tan $ , even at large depths. This favourable effect
of flexibility may however be balanced by an unfavourable effect due to
the extensibility of the inclusion, for soils whose residual strength is
very different from the peak strength.
Besides Texsol and reinforcement by continuous filaments, a relati-
vely important research has been performed and interesting results have
been published on reinforcement with fibers or sma~l plates. Although it
does not exhibit even a high resistance as Texsol, and has not yet given
rise to many applications, such a type of reinforcement appears to be an
interesting alternative, because of the easibility for mixing soil and
plates (or fibers) and the better isotropy of the reinforced soil thus
obtained.
The failure of a sand reinforced with small high strength inclusions
(metallic discs) was studied for the first time by Hausmann (1976) with
the triaxial apparatus. The failure obtained occurs by lack of adhesion,
and it is characterized by an apparent internal friction angle $a in-
creasing with the radius of the discs and by a cohesion value limited to
30 kPa when the size of the inclusions remains below 8 mm. These results
were confirmed by Verma and Char (1978) who carried out 120 triaxial
tests on a sand reinforced with small metallic inclusions having various
shapes.
The use of natural or synthetic flexible fibers, having low deforma-
tion moduli (Gray, 1978, 1983), or small pieces of ~eotextiles (Hoare,
1979 Mercer et al., 1984), show that the failure mode of the
reinforced sand by breakage of the inclusions is never obtained.
Figure 15 shows direct shear tests carried out by Gray (1983) on a
sand reinforced with natural fibers according to the procedure devised
by Jewell (1980). The fibers increase the shear strength of the soil,
but in different ways according to the angle they form with the horizon-
tal. The optimum improvement is obtained for a 60° angle corresponding
to the maximum extension line for the sand alone. A critical normal
stress defines two distinct zones for the failure curves; above, the
presence of the fibers results in a cohesion but it does not increase
the internal friction angle below, there is only an increase of the
internal friction angle.
 23

o
~200

00
00
w
a:
t-
OO o Non - reinforced
...J
« SLIPPAGE
.Y:f'
~
060"
~ 100
(l> x9(f>
z
~ A 120°
Fibers length : 4cm
Jd: max

200 300
NORMAL STRESS (kPo)

FIGURE 15. Shear strength of fiber reinforced soil (influence of the

fiber inclination) (after Gray, 1983)

This phenomenon can be explained as follows: in the first case, the

soil fails before the fibers because of the extensibility of the latter;
in the second case, there is a failure by lack of adhesion and slipping
of the fibers within the sand mass. This failure mechanism is different
from the one obtained for rigid inclusions in the triaxial apparatus and
for which the critical stress provides a distinction between a slippage
type of failure ~< G~ and a failure with reinforcement breakage ~ )0i)
(Schlosser and Long, 1972).
Thus, it appears that for a soil reinforced with fibers, breakage of
the fibers is practically never observed, except may be in the case of
very long and inextensible (metallic) fibers. The partial conclusion to
be drawn from this is that reinforcement with fibers improves the mate-
rial ductility and at the same time, it decreases dilatancy (Gray,
1983).
Research is at present time in progress, at Ecole Nationale des Ponts
et Chaussees, on sand reinforced with metallic or plastic fibers. Para-
meters such as fiber length, fiber extensibility, number of fibers per
volume unit, relative density of the sand are tested in the triaxial
apparatus.
It appears that a similar mechanical behaviour has to be expected for
a sand reinforced with plastic grid type square plates, 4 cm on side,
constituting very flexible inclusions (Mercer et al., 1984). Up to 0.6 %
in weight of inclusions, these plates do not increase the sand void
ratio, an? there is a 25 to 60 % increase in the maximum deviatoric
stress for a percentage of plates by weight of 0.19 %. However, due to a
large interlocking effect and entanglement between the inclusions and
the soil particles, noticeable shear strength increase occurs even at
low values of the confining stress. This type of soil improvement ap-
pears to be easy, cheap and effective.
24

Table II shows cohesion values for reinforced sand, obtained for dif-
ferent types of inclusions (fibers, small plates, continuous filaments),
for the same percentage of inclusion by weight (about 0.2 %). This table
also indicates the different soil-inclusion interactions involved.

INCLUSIONS FIBERS !GRID TYPE PLATES CONTINUOUS

5 cm !15 a 20 cm (4 x 4 cm) FILAMENTS

Cohesion !10 kPa 100 kPa 50 kPa 200 kPa

Types of !Friction !Friction !Friction

interaction !EXtensibility !Extensibility !Extensibility
! !Entanglement !Entanglement
!Interlocking !Curvature
! !effect

Table II Values of cohesion and soil-inclusion interaction types in

sand reinforced with synthetic inclusions

2.6 Another new retaining system: Tervoile

Among the--most recently developed retaining systems, it appears
interesting to mention Tervoile, which is both a frictional and
structural system.
As presented in Figure 16a, the basic structural element is a thin
corrugated and U-shaped membrane, (the corrugations being in a plane
perpendicular to the cylinder axis). This element is made of plane and
bent plates bolted together during construction. Until now, metallic
(galvanized corrugated steel) elements are used.
Figure 16b shows basic elements juxtaposed in order to form a retai-
ning structure. The needed interaction (i.e. pressure and frictional
shear stress) between the soil and the basic elements of the structure,
is provided by the backfill, spread and compacted in successive layers.
The structural elements work only in tension (or compression in the ver-
tical direction): this feature is characteristic ~f this particular
system.
Besides the backfill, the basic elements may either be used as the
only structural parts or be coupled with horizontal reinforcing layers
made of wiremesh, fabric membranes, etc .••
The mechanism of the retaining system is the following (Fig. 16c):
1) In a basic element, the backfill located in the bent part acts as
an active zone and applies pressure on the membrane. On the contrary,
backfill located between the two plane parts acts as a resistant zone.
There is some similarity with the reinforced soil system mechanism, but
the vertical location of the frictional elements and their large spacing
(2 m) makes the system very different.
2) The system has a relatively important rigidity in vertical planes
<

and can therefore withstand thrust by acting as a beam. This important

aspect of the mechanism enables a reduction of the length of the verti-
cal frictional elements.
 25

Corrugated
membrane

@ Basic element ® Tervoile retaining wall

I
I
,.,--,---------
/' I I
I I I
, I I
\ I
\ I
, .... I

Compressive

l
stress

I
I
I
I f+,~--!.'-''---'1 Thoust
\ i
' .... - ..I_-1---------
I

ACTIVE: RESISTANT
ZONE i ZONE

Anchoring effect Structural effect

® Behaviour mechanism

FIGURE 16. Tervoile retaining wall system

26

According to the inventor, V. Curt, a Canadian engineer, and his

french associate, A. Metulesco, this system is flexible and can tolerate
large differential settlements.
The advantages of such a retaining system are: 1) low cost, 2)
easibility of construction, 3) very light prefabricated elements.
A full scale experiment was performed in 1985 at Sherbrooke in Cana-
da, and in 1986 the Ministry of Transport in Quebec used this technique
in road construction at Grandes Piles for a retaining wall, 7 m high to
be built along a river.

2.7 Soil nailing

In-situ reinforcement of geotechnical materials by resistant in-
clusions is a very old process of construction often resorted to in the
past. Applications to in-situ retaining walls is more recent, although
the first reference in rock bolting is mentioned in 1930. In recent
decades, stabilization of rock slopes by passive metallic anchors grou-
ted within the rock mass has been intensively used. The purpose of this
technique is essentially to limit the decompression and the opening of
preexisting d~scontinuities by restraining the deformations, and to
create a mass in which rock blocks are locked together.
The first application of this technique to soils occurred in 1973
(Rabejact and Toudic, 1975) at Versailles (France), but the concept used
was different: the principle of soil nailing consists in obtaining a
homogeneous and resistant material by associating soil and inclusions
able to withstand tensile forces, but also shearing forces and bending
moments. However as indicated on figure 17, it can be considered that
soil nailed walls result from both rock slope stabilization and Reinfor-
ced Earth. The first soil nailed wall in Versailles was used as a
temporary support of the slope in order to build the definitive reinfor-
ced concrete wall, despite of difficult conditions (urban area, large
railway traffic).
Inclusions in a soil nailed wall are placed approximately horizontal
and are mainly subjected to tensile forces. They are generally made of
steel or another strong material like for instance, fiber-glass, and are
inserted into the soil either by simple driving or by grouting in a
predrilled borehole.
Figure 18a shows the different construction steps for a soil nailed
wall successive excavations with limited height (1 to 2 m) are
generally stable if the soil exhibits a short term cohesion, which is
commonly the case. Then the inclusions are placed. In most cases a
facing is required to ensure local stability between the bars it can
be made of shotcrete or steel mesh.
The behaviour of this type of retaining structure is very similar to
that of Reinforced Earth : the high elasticity modulus of the inclusions
restrains the displacements of the wall and modifies the shape of the
potential failure surface, which is different from the classical
Coulomb's wedge. This potential failure surface, which is the locus of
the maximum tensile forces in the bars, is parallel to the facing in the
upper part of the wall and divides the soil into two zones (Fig. 18b),
an ' active one and a resistant one. Moreover, the equivalent earth pres-
sure distribution related to the maximum tensile forces is quite diffe-
rent from that predicted by the Rankine's theory.
 27

Reinforced Earth Rock slope stabilisation

L_

~
e
v

Pragnieres wall (1965) Walls at Notre Dame de Commiers dam (1961)

(after Bonozzi et Colombet, 1984)

\ Soil nailing
/
",_ / Existing wall
I i'
, I: \I \
• I I

Versailles wall 1I973)

(after Rabejac et Toudic, 1975)

FIGURE 17. First soil nailed wall, reinforced earth wall and rock bol-
ted wall
28

1. excavation 2 . reinforced shotcrete

]
-~
. I
-
I
3. nailing 4. excavation

® SOIL NAILED CONSTRUCTION

Tmax

® SOIL NAILED BEHAVIOUR

FIGURE 18. Construction and mechanical behaviour of a soil nailed wall

 29

However, the behaviour of soil nailed walls is not yet very well
understood. In particular, the influence of the construction method
(excavation of steps) hasn't yet been investigated. Due to the great
interest of the technique, several research programs have been initiated
in different countries (Germany, France, USA). The french project CLOU-
TERRE on soil nailing is a 3 years research program : it mainly consists
of 5 to 6 full scale experiments, centrifuge model tests, laboratory and
field research on friction, in order to propose reliable design methods
including an evaluation of the inclusions durability. The aim of the
project is to promote the technique in order to use it for permanent
structures. Nail durability is therefore carefully considered an~ close
attention is given to nails specially designed for preventing metallic
rod or cable corros i on (figure 19).

Protective t ube with n

i jeclion holes

® " TBH A" NAIL PATE NTED AND DE VELOPED BY SOLRE NFOR

Steel cobles~'---~_.'1P"

@ PRESTRESSED MULTIREINFORCED NAIL "INTRAPAC" DEVELOPED

BY INTRAFOR - COFOR- (France)

FIGURE 19. Specia l nails deve loped for preventing corrosion

30

3. YAIN TYPES OF POLYMER SOIL RETAINING WALLS

3.1 Types of polymer used as reinforcements
Since the beginning of~einforced Earth development, tentative ef-
forts have been made for using polymeric reinforcements instead of
metallic ones. Compared to metals, polymeric materials have large ranges
of deformation modulus and tensile strength, and the following polymeric
products have been used as reinforcements: geotextile sheets, geogrid
sheets, woven geotextile strips, coated fiber strips, rigid plastic
strips. Figure 20 shows for instance some mechanical properties of
geotextiles and geogrids. Generally speaking, polymeric materials are
more deformable and less resistant than metals. Moreover, they exhibit
creep behaviour ; nevertheless it is possible to adapt in each retaining
system the type of polymeric reinforcement used to according the
allowable deformation.

TYPE MECHANICAL PROPERTIES

E Modulus Tensile
(secant at € = 100/ol resistance
(kN/m) (kN/m)

rn Non woven 2 -90 4- 35

w
..J
1=

•==}
><
w
f--
0
w
(,!)
Woven 50-1000 15- 350

§
c::::::J c=::::J
E
E
0
Q j 1 Extruded
"
og
In
C>
c:
c:
~
• • •

1
.,

p"",~
0

rn ;: 00000
c
'0
E
E 00000
. 00000 >
~ 0 50-700 9-90

·. . . . .
(,!) In
10
c:
0
W
0 00000
.,c:
"in 10
(!)

E
is

.
E
0
10 II}WoId'd bars
10

·.... ..
FIGURE 20. Types and mechanical properties of geotextiles and geogrids
 31

Inclusion extensibility greatly influences reinforced soil behaviour.

This has been very clearly shown by McGown et al. (1978), who have
considered extensible and inextensible inclusions. Besides increasing
strength, the principal action of extensible inclusions is to increase
soil ductility and decrease or even cancel the softening observed in
dense sand behaviour. Inversely, inextensible inclusions mainly increase
soil strength and deformation modulus, but they cause the deformation
soil modulus to be more brittle. These features are presented in figure
21 and allow the following distinctions to be made :
1 - Reinforcement with ideally inextensible inclusions, mainly repre-
sented by Reinforced Earth, for which the reinforcements are generally
linear and metallic.
2 - Reinforcement with ideally extensible inclusions, represented by
"p1y-soil" or "multi-layer soil" (McGown, 1978), for which the reinfor-
cements are generally plane and made of synthetic materials (geotex-
tiles, etc ..• ) .

-
..-- Sand wi th high strength
inextensible inclusion
Sand with high strength
_ _ _ _ _ .-/- extensible inclusion
b'
"- /
0
~ Sand with low strength
I inextensi ble inclusion
0
:;::
I
I /.~,
/ ' ... :"";':_"":"'::"":''';'':''':
~
II>
II> /
1/ ~--
",'/ Sand alone --,,' ~
- Dense sanl-.":

~
(i) 4
I Sand with low strength
l extensible inclusion

2 4 6 B 10
Axial slrain EI (%)

FIGURE 21. Deformability and strength inclusion influence o~ reinforced

dense sand behaviour (Mc Gown et al., 1978)

Table III gives a classification of reinforcement techniques as a

function of the relative deformability of the inclusion with respect to
the soil, and as a function of the geometrical type of design (one, two,
or three dimensional). Initially, and in particular for Reinforced
Earth, the reinforcement was made of linear inclusions, and thus it was
strongly anisotropic. Geotextiles and grids have led to the development
of two-dimensional reinforcement. Tridimensional reinforcement appears
theoretically isotropic, but it gives a high deformability to soils
reinforced , by a continuous filament (Texsol).
32

( Type of )
( In- rein. ! 1-D 2-D 3-D )
( clusion !
(
~~~--~~--------~----------~----------)
! )
( RIGID !Metal strips !Grids (metal- !Fibers (metal- )
( !lic) !lic) )
( )
( !Rigid plastic )
( !strips )
( ! ! )
( SEMI- !Grids (plastic)!Texsol )
( DEFORMABLE !Flexible plas- !Fibers (synthe-)
( !tic strips !tic) )
( ! )
( DEFOR..'1ABLE !Geotextiles )
(
----------~----------~----------~------~---)
Table III - Classification of reinforcement techniques according to the
inclusi~geometry and to its relative deformability

3.2 Polymer uses in Reinforced Earth Technique

Considering the whole research performed on Reinforced Earth,
particularly at the beginning in the 60's, it could appear surpr~s~ng
that nothing would have been done in order to use polymers as reinforce-
ments. In fact, Vidal (1966) planned to use at first polymeric
materials: " nylon strips, tergal strips and particularly rigid plastic
constituted of fiber glass coated with polyester resin. This last mate-
rial was chosen in 1965 and an important investment was made at that
time in order to produce industrially U shaped facing elements and
reinforcement strips. More specifically, this material was a fiberglass
reinforced plastic, in which strength and stiffness were imparted to
easily moulded resins by glass fibers. The individual glass fibers were
elastic and as strong as the strongest tensile steels, so that they gave
to the composite material a small deformability without creep and a high
strength. In 1965, this material had been used for 10 years in under-
ground pipelines and tanks and had proven to behave satisfactory. Howe-
ver, as indicated by Mallinde'r et al. (1977), some degradation was
observed when the material was maintained in wet conditions for long
periods of time.
In 1966, a first experimental Reinforced Earth wall using fibreglass
reinforced plastic strips and facing units was built, in order to test
the construction process and the mechanical behaviour of the wall.
Unfortunately, the plastic material was attacked by bacteria and the
Reinforced Earth wall was destroyed within 10 months. No biological test
was performed on the backfill material after failure and practically no
reliable information and expertise about the type of bacteria and the
degradability process is now available.
After Mallinder et al. (1977), this failure might have been acciden-
tal, since these authors gave biological test results on fiberglass
reinforced plastic, indicating that this type of material was not
degraded by bacteria (immersion time was however limited to 6 months).
Nevertheless this failure has been the turning point for the use of
plastic materials in Reinforced Earth: at the beginning of 1967, VIDAL
decided to develop Reinforced Earth with metallic reinforcements.
However, the use of plastics in Reinforced Earth was not yet
completely abandoned, since another very interesting attempt was done in
 33

1971 with the construction of the Poitiers wall using Tergal strips.
This wall was a temporary structure, 5 m high and 40 m long. The tergal
strips were attached to cruciform concrete panels. The calculation of
the wall took into account tergal creep and during its service life, the
wall behaved satisfactory. However, it appeared during construction that
tergal strips had to be slightly prestressed in order to prevent exces-
sive lateral displacements of the facing.
In 1981, ten years after its construction, the wall was dismantled
and interesting durability tests were performed on the strip material.
As presented further, it was shown that the plastic fibers had been
degraded and that the mechanical properties of the plastic strips had
decreased.

3.3 Geotextile retaining walls

As previously indicated, the first reinforced soil wall using
geotextile was built in 1971. The backfill material used was a wet
clayey soil and the foundation soil was poor and compressible (peat).
It appeared from this first encouraging construction that, compared
to Reinforced Earth type walls using strip reinforcements and _ facing
units, a geotextile wall presents the following features :
1) It is a very cheap retaining system.
2) Geotextiles enable the use of poor backfill material, because of
their large frictional surface area when they are two-dimensionnaly
used, and because of their drainage properties.
3) The most important problem is the control of the displacements an.d
the deformations, particularly when using non-woven geotextiles.
4) Building vertical walls is rather tricky.
5) The facing is generally rather bad-looking.
Despite of advantages (1) and (2), such a retaining wall system has
been little used until now. Concerning the construction of vertical
walls, Jones (1977) has compared different methods which are presented
in figure 22. For the first walls built in France (Puig et al . , 1977,
Kern, 1977), temporary lateral supports were used: embankment or wood
plate. It appeared further that a good way of controlling the verti-
cality and the facing deformations was to use posts in addition to the
facing geotextile (Jones, 1977; Schwantes, 1982). But some other
methods are also suitable for controlling lateral displacements and
deformations, furthermore leading to aesthetic f a cings. Ihey are es-
sentially of two types (Figure 23): 1) use of prefabricated vertical
facing units (concrete panels) (Broms, 1977; Jones, 1982); 2) cons-
truction of an inclined facing using another type of reinforcement
(grid, Armater, Texsol). Delmas et al. (1986) have published an interes-
ting paper dealing with these different facing construction methods.
The low cost of all types of geotextile reinforced soil walls is thus
unfortunately counter-balanced by large deformations and inaesthetic
aspect of the facing. In order to keep the cost advantage and to avoid
facing deformations, mixed structures have been developed :
1) walls with prefabricated or built in place facing units (concrete
panels, gabions, etc •.. ) and geotextile reinforcements (Broms, 1977 ;
Jones, 1982; Delmas et al., 1986).
2) Walls with fabric facing and relatively inextensible
reinforcements (metallic strips , geogrids) (Schwantes, 1982).
It seems that such combinations could lead to interesting
developments in the future.
34

® U.S. Method

® French method

Corner detai I

~ Detail
showing ~
.....,
.:
"'; ':
':....
':".!: ......:
.

A
rod
settlement
M: .. ..
......... ......,,:
:-
"

Construction
sequence

® British Yorkshire method

FIGURE 22. Comparison of vertical geotextile wall construction methods

(J o!1es, 1977)
 35

q
HHHLH

Gobion filled with r'---"

r
compacted sand

t--'~=r-~~~++:~'
I~~'--'" 'h~
L~--r'"--..r-~~---r-~---r---r--.J::..
@ Concrete facing units ® Gobion facing units

Geogrld

Armater sheet

Moltimum
slope

@ Inclined facing with geogrid @ Inclined facing with "Armoter"

FIGURE 23. Different types of improving geotextile wall facings (Delmas

et al., 1986)
36

Because of their relatively large extensibility, no deformation and

tensile stress measurements have been carried out in the geotextile
reinforcement sheets of actual walls. However it is now well recognized
that the behaviour of a wall reinforced with geotextiles is controlled
by the reinforcement deformations and it is therefore different from the
behaviour of a wall built with quasi inextensible reinforcements (Rein-
forced Earth). This fact results from theoretical considerations
(McGown, 1978) and from triaxial tests on sand samples reinforced with
geotextiles (Blivet, 1979; Gray et al., 1982). In particular, although
reinforcements with very extensible synthetic fabric increase the ulti-
mate strength, they tend to reduce the overall stiffness of the sand.
Delmas et al. (1986 b) have recently developed a design method for
geotextile soil reinforced wall. Contrary to the limit equilibrium me-
thods, this method is based on an evaluation of the deformations and
comprises the three following steps for calculating the safety factor
along any potential failure surface :
1) Determination of a displacement field compatible with the accepta-
ble wall displacements at the facing and at the top.
2) Determination of the stresses in the geotextile sheets resulting
from the above displacements.
3) Checking of the reinforced soil mass equilibrium along potential
failure surfaces.

3.4 The York method

The York method for the design of reinforced soil retaining walls has
been developed by. JONES (1973) on behalf of the Departement of
Transportation in U.K. This method is similar to the Reinforced Earth
technique despite two small differences :
1) All the prefabricated elements (facing units, strips) can be made
of plastic.
2) Reinforcing strips can slide with respect to the facing thanks to
the use of vertical poles.
According to JONES (1977), differential settlements can easily be
accomodated by the York method, also called the sliding method.
Figure 24 shows the comparison between Reinforced Earth and York
method prefabricated elements. In the York method, the facing units are
hexagonal membranes, made from glassfiber reinforced cement and bolted
together. Strips are either metallic or plastic (fiherglass reinforced
plastic as previously described).'
Until now, this type of polymer reinforced soil wall has only been
developed in U.K. and even in this country a few number of walls have
been built. Although nothing has been published yet about the long term
behaviour of these structures, it seems that plastic membrane facing
units are brittle and could have been damaged. In fact, a relatively
long experience in reinforced soil wall construction shows that concrete
panel is one of the best and the most resistant facing units. Despite of
this, the York method has to be remembered as the first reinforced soil
wall totally built with plastic material.

3.5 Geogrids reinforced soil retaining walls

The satisfactory behaviour of the Dunsmuir wall (1974), in which
welded metallic bar-mats were used, and subsequent research performed o~
soil-grid friction have proved that grids were suitable for reinforcing
a large range of soils.
 37

Temporary
wedges Reinforcement

Temporary
clamp

@ Reinforced Earth (telescope method of construction)

@ York method (slid inO method of construction)

FIGURE 24. Comparison between reinforced earth and York method

38

Furthermore, t~e relatively large development of geogrids at the

beginning of the 80's led to an attempt to use such reinforcements in
retaining wall systems. At present time, only a few walls have been
~onstructed with geogrids and, it seems, because of the same facing
?roblems than those encountered with geotextile walls. Considering the
interest presented by geogrids in soil reinforcement, the use of
~refabricatp.d facine units associated with an easy and efficient
attachment system would certainly be an interesting possibility of
development.
However, special attention has to be given to creep and durability.

3.6 Websol
This type of soil retaining system has been developed in U.K. by Soil
Structures Limited at the end of the 70s and it is very similar to the
Reinforced Earth technique.
Paraweb is a composite plastic material consisting of polyester yarns
(terylene) coated and protected by a sheet of black polyethylene
(alkathene). The fibers give high strength as well as relatively low
creep characteristics to the material, and the Alkathene sheet protects
it from ultraviolet rays as well as other chemical degradation. Despite
its extensibflity, Paraweb appears to be an interesting plastic material
for soil reinforcement. The Paraweb strips are 2 mm thick and 90 mm wide
(figure 2Sa), and its rupture load ranges between SO and 100 kN.
Facing elements are T-shaped concrete panels (figure 2Sb). The rein-
forcements are formed by a continuous strip which has to be prestressed,
because of the extensibility of Paraweb, in order to avoid displacement
of the facing (figure 2Sc).
John et al. (1983) have performed two full-scale experiments on
Websol reinforced soil walls. The distribution of maximum tensile forces
measured the one corresponding to inextensible reinforcements as steel
strips. In the upper part of the wall, it appears to be close to the Ka
line, suggesting that the lateral deformations of the structure are
large enough to attain the Ka state of stress in the soil. This effect
of the extensibility of the reinforcement strips on the maximum tensile
forces has since been confirmed by several studies ; for instance,
fiberglass reinforced plastic behaves as an inextensible reinforcement
material. Reinforcement extens·ibility has also an influence on the
potential failure line (maximum tensile forces line) ' which is close to
Rankine's plane in the case of extensible reinforcements.

3.7 Texsol walls

Since its first use at Caudebec-en-Caux (Leflaive et al., 1983),
Texsol has been until now essentially used in retaining wall
construction.
Leflaive and Liausu (1986) presented a large project in which Texsol
was used for enlarging the french A7 highway. Figure 26a shows a compa-
rison between two possible solutions: a concrete wall and Texsol.
Texsol was chosen because of the following requirements
- Construction had to be achieved very quickly.
- Environment protection was to be considered.
- Traffic had to be maintained.
- Low cost.
It must be noted that the Texsol walls were grassed to provide good
environemental conditions and facing protection.
 39

L(JoN lEVEL TYPE

----------\ I

I
\

DETY'
( Q"i\'~"~::E:=S(f«"'THYlE"'1
TER't'LENE (POLYESTER)

® Para web strip ® Websol wall facing

Precast concrete facing panel

CD Plan view of the strip placing

FIGURE 25. Websol retaining wall system

40

CLASSICAL SOLUTION
WITH CONCRETE WALL
} TEXSOL WALL
SOLUTION

i
Natural slope
i,
Concrete
I,
wall I
profile

® Comparison between classical solution and Texsol solution

1.51

SOIL

rc::ol
Highway ~
o90m.1 L
® Texsol wall design method

TEXSOL COHESION SAND INTERNAL FRICTION

(kPo) ANGLE '</>'= 380
300

200

100

THREAD CONTENT
BY WEIGHT (%)

0.5 1.5 2 2.5 3

(£) Texsol cohesion value versus thread content by weight

FIGURE 26. Texsol wall solution used for enlarging the french A7 highway
(1985-1986)
 41

The following geometry was adopted for the walls :

- External facing inclination: 60° on the horizontal plane.
- 0.5 m base embedment in the foundation soil.
- Height ranging from 3 m to 7 m.
- Width at the top : 0,55 m.
- Width at the base : determined from design calculations in order to
have a minimum slope/wall stability of 1.5 .
As mentioned by Leflaive and Liausu, there are two steps in Texsol
wall design :
1) Considering the available granular material, triaxial tests have
to be carried out on reinforced soil samples in order to determine the
influence on Texsol cohesion of parameters such as thread resistance and
thread content by weight.
2) Wall stability along circular potential surfaces has to be ana-
lysed and wall geometry must be adapted in order to obtain a min~mum
safety factor of 1.5 .
This wall design must be considered as a first approach and it will
certainly need to be further improved in order to account for the two
following phenomena related to Texsol behaviour. The first one is the
cohesion anisotropy resulting from placement technique used for Texsol :
Texsol cohesion determined from triaxial tests may thus not be complete-
ly representative of the actual cohesion mobilized along horizontal
planes. The second phenomenon is the difference in the at peak deforma-
tions observed on the stress-strain curves of soil and Texsol. This
difference certainly modifies the failure criterium to be taken into
account along the potential failure surface.
Placing Texsol is carried out with a large special machine which
builds successive horizontal layers 5 to 10 cm high. Compaction of
Texsol is done by a vibrating plate, in the lower part of the wall
(up to a height of 1.6 m above the base).
On the A7 highway project 20 000 m3 of Texsol have been used for
the construction of retaining walls and behave satisfactory.
Texsol has promoted the use of three-dimensional soil reinforcement,
which certainly will be developed in the future not only for retaining-
walls, but in other fields like foundations where such a reinforcement
appears to be particularly attractive. Fibers and mesh elements must be
considered as potentially interesting reinforcements.
However, it seems that future development of Texsol req~ires further
studies that should be focussed on the following points
- Mechanical behaviour of the material with respect to the placement
method, the deformations, etc ...
- Observations on real structures and for instance on centrifugal
models brought to failure, in order to improve the design methods.
- Durability of the material with respect to the facing protection
and creep.

4. PROPERTIES OF POLYMERS WITH RESPECT TO THE BEHAVIOUR OF REINFORCED

SOIL WALLS
When metallic reinforcements are inserted into the soil, they may be
considered as rigid with respect to the soil deformability, and to the
magnitude of the induced stresses. On the contrary, polymeric inclusions
are characterized by weaker mechanical properties, i.e high exten-
sibility, low'tensile strength, associated with long term creep. Fur-
thermore, although not sensitive to electro-chemical corrosion, polymers
may also suffer some degradations under soil physico-chemical environ-
42

ment, and durability problems are to be considered.

Several studies have been conducted on strength-strain properties of
geotextiles since they have been used in various applications of
geotechnical engineering. It has been observed that among all those ap-
plications, reinforced soil retaining walls constitute a case in which
the magnitude of stresses applied by the soil to the geotextile is the
higher. In fact, it is quite difficult to fully understand the load
transfer mechanism which occurs between the soil and the textile, and to
know the exact stress field the textile is submitted to.

4.1 Extensibility
4.1.1 Polymers properties
Table IV shows some typical mechanical data for the more current po-
lymers used in geotechnical engineering, as compared to steel proper-
ties. It should be noticed that, for each polymer, both bulk material
and filament properties are given. In fact, filaments are produced by
extrusion of the heated polymer mass, and high tenacity yarns, as des-
cribed in the table, are obtained by controlling the cooling process of
the filament, and by stretching it during extrusion. In such a way,
polymeric chains have a preferential orientation and an increased aniso-
tropy, thus resulting in much higher (up to ten times) mechanical
properties of the material, such as tensile strength or elastic modulus.

(
( Elastic modulus !Tensile strength !Failure elonga)
( ! ! (MFa) (MFa) !tion filament)
( Material !Density! (%) )
( ! ! ! ! )
( ! Bulk !Filament! Bulk !Filament )
( !material! !material! )
( )
( )
(Steel 7,85 200 000 2340 3 )
( )
(Polyester 1,38 2100 13500 60 '800 6-20 )
( 18500 )
( )
(Polypropylene 0,91 1100 7400 35 !380-780 15-25 )
( )
(Polyamide ••• 1,1 2400 12500 40-120 !600-900 12-26 )
( )
(Polyethylene 0,95 800 4300 30 !390-700 10-20 )
(high density 6500 )
( )

Table IV - Typical polymers characteristics, as compared to steel

It may be seen from table IV that the main difference existing

between metallic and polymeric materials corresponds to the high exten-
sibility of polymers. This is illustrated by the values of elastic
 43

Tensile stress
(MPo)

2800

2400 /Steel thread

.'F'Iberglass

2000

1600

1200 /Polyester
Nylon

800

400

Extelllion °4

5 10 15 20

FIGURE 27. Stress-strain behaviour of different material fibers

44

moduli, which are, for current polymers, from 10 to 30 times lower than
for metals ; the tensile strength is also lower (2,5 to 5 times),
whereas the deformation at failure is much higher (20-30 % instead of
3 %). When comparing current polymers, it may be seen that polyester
and polyamide, which have higher densities (1,38 and 1,1 respectively)
have higher stiffness and tensile strength than polypropylene and polye-
thylene high density (0,91 and 0,95 respectively).
Figure 27 shows tensile test results for steel, fiberglass, polyester
and polyamide filaments. Whereas failure occurs, for steel, at a defor-
mation of 3.2 % for a tensile strength of 2340 I1Pa, polyester and nylon
fail for much larger deformation (respectively 11 % and 19 % on the
figure), and a tensile strength of about 1000 Mpa.
Baudonnel et al. (1982) took some filaments from current polyester
and polypropylene geotextiles, and elongation tests results are shown in
figure 28.

F
A (N hex)

0,3

)(

~
~
..-< 0,2 PoJypropylene
en
en
0
E
...
0
CI>
c: Polyester
SD A =7.35
LT A =2.45 tJ. • (dtex)
-I
"-
Polypropylene
SD A = 7.1 •,
u. TP A =12.8 0

CI> 0,1-
...
v
~

oL------------,--------____.-__________-.______ _
£ (%)

° 50 100 150
Strain

FIGURE 28. Elongation tests on filament extracted from non-woven fabrics

(Baudonnel et al., 1987)
 45

The strength value is expressed in terms of tenacity, say the force

divided by the linear mass of the yarn, expressed in N/tex. In the case
of non-woven fabric extracted filaments, the deformation at failure is
in both cases quite higher (71-78 % for polyester, and 155-239 % for
polypropylene). The polypropylene curve is composed of two parts having
different slopes, the first one being identical to the polyester one.
These results illustrate the decrease of the mechanical properties which
occur when fibers are processed for fabric construction.
The previous data were related to solids and fibers, but the great
variety of polymeric inclusions used in geotechnical engineering may
exhibit various properties, which do not only depend on the nature of
the polymer, but also on the structure of the inclusion, and on the
influence of the soil confinement.
The influence of those parameters has been explicitely evidenced by
McGown et al. (1982), who performed load-extension tests on geotextiles
confined in soil, with the help of the apparatus presented in figure 29.

Dial
gouge

Top clamp

Reinforced
geotextile
Pressure bellows
LH' __-M--.... · Lu bricated
Geotextile --l,;,t----{!l membranes

Lubricated __ -",..--,., [jIO-Ii;t--- Ru bber

membranes
Confining
Reinforced
geotextile sand

Bottom clamp
1~f'1J~~,,¥'------ Bottom
clamp

FIGURE 29. In soil tension apparatus (Mc Gown et al., 1982)

46

In this apparatus, a geotextile is included between two layers of soil

(Leighton Buzzard sand, Dso = 0,85 mm), and pressure is applied on each
side of the fabric by two air-activated rubber pressure bellows. McGown
et al. tested four different fabrics : woven, non-woven meltbonded, non-
woven needle punched, composite woven and needle punched. The elongation
tests were performed at 20 c e and at a constant rate of strain of 2 % per
minute. Some of the results obtained are presented in figure 29. It may
be seen that for the woven polypropylene fabric (Lotrak 16/15, 120
g/m~, the in-soil confinement has a very little influence. On the
contrary, in the case of a polyester non-woven needle punched fabric
(Bidim U24, 210 g/m 2) , the effect of in-soil confinement is quite
important, and a 100 kPa confinement pressure results in a strengthening
of the fabric, which corresponds to a higher stiffness, as illustrated
by a higher slope of the elongation curve. This phenomenon may be inter-
pretated in the following manner : the fabric elongation does not invol-
ve individual filaments, but rather induces a rearrangement of the
needle punched filaments, which affects the bonds between the filaments.
When soil is in contact with the fabric under a given pressure, it
contributes to the stability of the bonds between the filaments, and
provides a higher strength to the fabric.
14r------r------r------.----~

12

10

",.
/'
8 /
e /
"- /
...
z
/
/
/

0 /
« 6 /
9 /
/
Non
/4:::0,.. woven
/ polyester
I
4 /
/
I
/
/

,/
/

/
- - - In(IOOkN/m
soil confined
3)
/
/
/ ------Unconfined
/
/
0
0 10 20 30 40
AXIAL STRAIN (%)

FIGURE 30. Influence of the in-soil confinement and of the fabric struc-
ture on elongation properties (Mc Gown et al., 1982)
 47

It is interesting to notice, on figure 30, the predominant influence

of the structure of the fabric as compared to the polymer type, since a
120 glm 2 polypropylene woven fabric is stiffer than a confined 210 glm 2
non-woven needle punched polyester fabric. For the woven fabric, failure
occurs at 23-28 % deformation, whereas for the non-woven material, lower
levels of load induce deformations as high as 40 %, without occasioning
rupture.
The problem of tensile testing of fabrics is then particularly im-
portant for non-woven fabrics, since unconfined testing induce an impor-
tant retraction of the strip, as shown in figure 31. For this reason, an
initially current textile test which consisted of stretching up to
failure a 50 mm wide and 200-300 mm long strip has been replaced by
some other tests, with wider strips.

t Load

I
I
I
I
I
I
I

:E
I

FIGURE 31. Lateral contraction occuring during an elongation test on a

non-woven fabric

Presently, agreement does not exist on the best width to be selected

for the strip. McGown et al. (1982) have studied the influence of the
width on unconfined elongation tests and have compared their results
with in-soil 200 mm wide tests results (figure 32). It may be observed
that, due to retraction, width has a great influence on strength-strain
behaviour, and that strip widths of 50 or 100 mm are definitly too
small. It is interesting to notice that the 200 mm test corresponds to
the 0 kPa in-soil confined test, whereas the 500 mm test corresponds to
the 10 kPa in-soil confined stress.
In fact, present discussions on strip width are related to those two
values of 200 and 500 mm, for a length of 100 mm, and argumentation
elements have been obtained from special tests, where lateral retraction
was avoided. Such special tests include a hydraulic tensile test
(Raumann, 1979), biaxial tensile tests' (Viergever et al., 1979), cylin-
drical sleeve test (Paute et Segouin, 1977), and a special test in which
lateral restraint is achieved by means of lightweight wooden brackets in
which steel pins have been set (Sissons, 1977). The fabric is pressed on
ten of these brackets regularly scattered along the length of the 200 mm
48

wide strip. The pins cross the fabrics and avoid restraint. Shresta and
Bell (1982) have used this device for testing several geotextiles.
Testing of 200 mID wide strips has been also performed by Murray et al.
(1986), and Richard and Scott (1986).

12r-------------------------~
-----Unconfined in Isolation

10

100
E 8 55
"-
...z 10
/1'"",
c o ,
~
g 6 ",,"
"
/Size of
/ test
, specimens
4
(mm)

400)( 100
2 / 300)( 100
/"~ ~ 200 )( 00
, ,," ~ 100 )( 100
....... ........

10 20 40
AXIAL STRAIN !%)

FIGURE 32. Unconfined in-isolation and confined in-soil load-axial

strain data for Bidim U24 (Mc Gown et al., 1982)

Leflaive et al. (1982) performed 500 mID wide tests, and proposed a
correction corresponding to the lateral contraction. The corrected
results they obtained compared favourably with results of cylindrical
sleeve test, and in-soil confined tests (McGown et al., 1982). This
approach has been adopted by Cazuffi et al. (1986), Leclerq and Prudon
(1986). Rowe and Ho (1986) also suggest a 500 mID wide value.
Other parameters have been studied. Figure 33 shows the influence of
the strain rate on the maximum tensile strength of some woven fabrics
(Rowe a nd Ho, 1986). In this case, where the constitutive fibers are
directly sollicitated by the tensile test, the effect of strain rate is
import a nt, and the writers suggest a 2 % r a te. Several writers conside-
red the influence of the tensile direction, as compared to warp, weft,
or diagonal direction for woven fabrics, and production direction for
non-woven fabrics. In the later case, Van Leeuven (1977), Leclerq and
 49

E
......
...
z 160

:I:
I-
(!)
Z
W
II::
I- 120
If)

w
...J
iii
z
W
I-
::I 80
::l
::I
x
<t
::I

o
40
-r
- - i ]~------------~o~--------~o----

0
0.1 10 100
STRAIN RATE( %/mln) - LOG SCALE

FIGURE 33. Variation in tensile strength of some woven fabrics with

strain rate (Rowe and Ho, 1986)

Prudon (1986) observed no variations of the tensile strength, whereas

Paute et Segouin (1977) mention some decrease in the production direc-
tion. In the former case, warp and weft directions give generally simi-
lar results, whereas a 20-40 % decrease is observed in diagonal direc-
tions. However, Rowe and Ho (1986) observed significant variations on
some woven fabrics, the warp direction being sometimes stronger.
For non-woven fabrics, and for a given type of polymer, the tensile
properties are dependent on the weight per unit area of the fabric, as
shown on figure 34 (Paute and Segouin, 1977), from results of the cylin-
drical sleeve. The curves show, for various non-woven polyester and
polypropylene fabrics, the influence of the value of the weight per unit
area, expressed in g/m 2 • The mechanical data are the rupture strength
(34a), and the deformation modulus E (34b). The increase at rupture
strength is fairly linear for all fabrics, whereas the deformation
modulus shows, in the case of polyester, a slope increase of about 400
g/m 2 • It is interesting to notice that the nature of the polymer has
50

little influence on the mechanical properties of the material. These

curves give an idea on non-woven rupture strength, which may vary,
according to the weight per unit surface, between 10 and 50 kN/m whereas
the deformation modulus is comprised between 30 and 200 kN/m except for
the weaker fabric.

50~-------------+-+~ 200r-------------r-+_

e
.......
z
~ 40r-------------~~-- e
.......
C
z 150r--------------i~--
~
<I:
9 --- Polyester en
::J
~ 30 ~----------j>---/'---+---­ --- Polypropylene ..J

...
::J
II.
::J
C
~ 1001--~~-----.~+-+----
::J
II:
z
20r-----~~L----+--- Q
~
2
II:
~ 50r---~~------~~---
1£1
10r---'---+--------+--- c
I
'Y
/

00 500
SURFACE WEIGHT (gl m2 ) SURFACE WEIGHT (g/m 2 )

FIGURE 34. Tensile strength and deformation moduli of polyesters and po-
lypropylene fabrics listed on the sleeve-cylinder apparatus (Paute and
Segouin, 1977)

For woven fabrics, tensile strength data depends more on the

individual filament characteristics. Table V shows data from Roscoe and
Ho (1986).

Woven !Initial tangent Tangent modulus

Geotextile modulus (kN/m) from 10 % strain
(kN/m)

Geolon 1250 200 800

Permealiner 1195 220 140

Table ~ - Properties of two woven fabrics (Rowe and Ho, 1986)

 51

For geogrids, the secant modulus at 10 % strain is included between

50 and 700 kN/m, whereas the tensile resistance varies between 9 and 90
kN/m. It must be noted that, as mentionned previously, extruded geogrids
have better mechanical properties that punched ones.
The table of figure 20 gives compared values of E Modulus (secant at
10 %) and tensile resistance of the various types of polymers used in
reinforcement. It may be seen that, due to stretching of the filaments,
woven geotextiles exhibits the better properties.

4.2 Creep behaviour of polymers

Iu--the case of metallic reinforcement, the level of stress induced
through soil-structure interaction in retaining structures is quite low
as compared to metal tensile strength, and creep of metals is not signi-
ficant. On the contrary, the tensile strength of polymers is much
lower, and creep behaviour has to be seriously considered for assessment
of long term stability of retaining structures.
Creep of polymers yarns is a well known phenomenon, and figure 35
presents results of long term elongation tests, on yarns submitted to
constant tensile loads (Greenwood and Myles, 1982), during 10 000 hours,
i-e 1 year and 50 days. For this duration, and for loads not exceeding
40 % of the rupture load, strain is a linear function of the logarithm
of time, as in many other materials like soils, for example. Polyester
yarns are characterized by a relatively high instantaneous strain. The
values of the instantaneous strain and of the creep rates are in good
agreement, for polyester, with previous results presented by Finnigan
(1977). As compared to polyester, polypropylene exhibits lower instanta-
neous strain, but higher creep rates. This creep tendency of polypropy-
lene is well known among textile people. At 60 % of the rupture load,
there is an upturn of the curves, which may be characteristic of the
rupture phenomenon initiation.

TOTAL STRAIN (%)

_ _ _ Polyester

- - - - Polypropylene

~~~~
~~
----------
l--~__~==~~~~~::==------_:~~~~--~~~=-~----------------
-- -- ---
~~
5 .... ----

-~
---------------

0.1
-- -- -- --- 10 100
TIME
1000
109 t (hours)
10000

FIGURE 35. Creep of polyester and polypropylene yarns (Greenwood and

Myles, 1986)
52

In fact, long duration tests mentionned by Greenwood and Myles on

other polymers yarns (parafilrape) during seven years, showed such
upturns, which were initiated after more than 10 000 hours testing.
Therefore, linear extrapolation of the curves for long durations may not
be realistic, and it may underestimate creep strains. As mentionned, by
Finnigan (1977), techniques such as heat stretching of the filament (12 %
to 10 % at 235°C for 75 seconds) may significantly reduce the creep
tendency.
In the case of woven fabrics, data from Van Leeuwen (1977), presented
in the discussion session of the Paris Conference (Vol. III, p. 102) are
shmm in figure 36.

£{%l

20
~-
~ Q{oQ'i
?o:'1 ____
- I I

----
16
~ Polyoml'd e -----
12
Polyester -----

T
8 ~ I-~
I
i

1
4
t {days I
o
0.001 QOI QI 10 100 1000
I
,
I

I
I min I hour I day

FIGURE 36. Creep of synthetic woven fabrics under prolonged loading (50%
of the breaking strength) (Van Leeuwen, 1977)

These data concern polyester, polyamide and polypropylene, for loads

equal to 50 % of the rupture strength. As for short term elongation
tests described previously, the properties of the constitutive filament
has a strong influence on the behaviour of the woven fabric. The best
creep behaviour is observed for polyester fabric. Polyamide fabric
exhibits a little higher tendency to creep, whereas the creep observed
for polypropylene fabric is quite important. However, the author men-
tions that no heat treatment was performed to improve creep properties
of those polymers fabrics. In the case of polyester, table VI shows a
comparison between the creep of yarn and woven fabrics, based on strain
values corresponding to one decimal logarithm cycle (creep rate).
It may be seen that few differences exist, in terms of creep
behaviour, between yarns and woven fabrics. Such a statement has former-
ly'been made by Finnigan (1977).
Such a comparison made for polypropylene woven fabrics shows, accor-
ding to the results of Van Leeuwen, that a larger increase in creep is
observed when passing from yarns to fabric; creep rates at 50 % of
breaking load increase from approximately 1,13 (Greenwood and Myles) to
2,1 (Van Leeuwen). Bell et al. (1982) mention the possibility of impro-
 53

ving the polypropylene yarns creep rate under 40 % breaking load, from
1,5 to 0,40 by an adequate treatment. However, for woven fabrics, they
only give creep rates at 20 % breaking load (0,40 - 0,73), which does
not allow a direct comparison to be made.

)
Reference Type of % of breaking Creep rate )
material load ! (% extension for )
!per log10t cycle»
! )
)
(Finnigan (1977) ! Polyester yarn! 41,2 0,192 )
( ! ! )
(Finnigan (1977)!Polyester yarn! 58,8 0,194 )
( ! )
(Greenwood and )
(Myles (1986) !Polyester yarn! 40 0,125 )
( ! ! )
(Greenwood and )
(Myles (1986) !Polyester yarn! 60 0,27 )
( ! ! )
(Van Leeuwen Polyester )
«(1977) woven fabric 50 0,2 )
( )

Table VI - Comparison of creep properties of polyester yarns and woven

fabrics

For the SR2 Tensar geogrid, composed of high density polyethylene,

data from the isochronous load-strain curves presented by McGown et al.
(1984) are reported in strain/logarithm of time diagrams in figure 37.
TOTAL STRAIN
(%)

10

DURATION, HOURS
OL--------.---------.--------.--------.----~--_,--~
0.1 10 100 1000 10000
FIGURE 37. Creep behaviour of tensar SR2 (after Mc Gown et al., 1982)
54

It is interesting to note that the creep rates of those polyethylene

grids are in the same order of magnitude as the polypropylene yarns
studied by Greenwood and Myles (1986). Unfortunately, such data were not
found for the polypropylene SS2 geogrid, which however should exhibit a
similar behaviour. In fact, due to the structure of the grids, it seems
logical to consider that, as in the case of woven fabrics, geogrids
should have a creep behaviour directly related to their composition.
As in the case of elongation tests, problems arose for investigating
the creep behaviour of non-woven fabrics, because of the lateral
contraction of the · elongated strip during non confined tests. For this
type of test, predominant influence of the structure of the fabric and
of the soil confinement was evidenced, as shown in figure 38 (McGown et
al., 1982). The creep rate decrease is particularly evident in the case
of Terram 1000 (67 % polypropylene; 33 % polyethylene melt bonded). For
Bidim U24 (polyester, needle punched), the reduction is more efficient
in terms of instantaneous strain. Such in-soil creep tests are not yet
very numerous, and discussion on the combined effect of polymer composi-
tion and fabric construction still has to rely on non-confined tests.

TIME (hours) TIME (hours)

0·01 0·1 1·0 10 100 1000 0.01 0·1 10 10 100 1000

l-
Unconfined I
30 I
In-Isolation ~ I

16
------------ -,... 40 Unconfin .. d I
I

In -Isolal ion-,-
20 /
/
/
;?12 Confining pressure
z 30 ,,/' Confining
z 100 kN/m2
...
~
cc ,"" pressure
,10------
-
4 /
~ IF> 100 kN/m2
...cc 8
IF> ~20 I
>< I
<{ I
I
I
I
10 I
/
/- '\.\.'+-~
, -i>
-s>
Irerram 1000 I 0<::.
Bidim U24

FIGURE 38. Influence of in-soil confinement on creep properties of two

non-woven fabrics (Mc Gown et al., 1982)

Table VII shows the results of creep tests which were performed on
200 mm wide strips by Shrestha and Bell (1982). Lowest creep was exhi-
bited by polyester resin bonded NW1, and highest creep by polypropylene
staple needle punched NW6. Creep is most sensitive to load levels for
the continuous filament polypropylene geotextile NW3 (Heat bonded) and
NW5 (Needle punched), which show a 3 to 5 fold increase in creep when
the sustained load is doubled. The similar results obtained for these
two fabrics indicate an identical effect of heat bonding and needle
punching regarding creep behaviour. Staple filaments (NW6) seem quite
unfavourable with respect to creep properties. For NW1 and NW6, creep
increased by only 1 % to 2 % for increases in the sustained load levels
 55

ranging from 40 % to 57 %, and 33 % to 56 % respectively. The creep of

polypropylene woven fabrics W4 and C1 is higher than the creep of non-
woven polyester fabric NW1, and smaller than the creep of all non-woven
polypropylene fabrics.

Creep measured in 20 hours

(Fabric! Geotextile Filament
( construction !
( !Load level!Creep!Load level!Creep)
( (%) (%) (%)! (%) )
(----~---------7---------7------_T--~------~----))
( ! !
(NW1 !Non woven !Polyester 40 3 57 4)
( !resin bonded !continuous )
( )
(NW3 !Non woven !Polypropylene! 35 5 63 27)
( ! hea t bonded ! continuous ! )
( ! )
(NW5 !Non woven !Polypropylene! 33 9 57 31)
( !needle punched!continuous ! )
( ! ! )
(NW6 !Non woven !Polypropylene! 33 20 56 22)
( !needle punched!staple )
( ! )
( W4 !Woven ! Polypropylene! 31 11 44 16)
( !mono filament! .! )
( ! )
( C1 ! Woven, with ! Polypropylene! 36 5 55 8)
( !needle nap !slit film )
( )
( )

Table VII : Creep properties of various fabrics

Elements on the relative influence of both polymer composition and

geotextile construction may be drawn from results of Allen et al.
(1982), as shown in figure 39.
56

CREEP STRAIN
(%)
250

200

150

300
100

5O."--_~

TIME (minutes)
O~-----'-------.------.-------.------.-------.-------r--~
0.1 10 100 1000 10000 100000 1000000

CREEP STRAIN
(%)
50

POLYPROPYLENE

J-
40

Typor 3401
30

• .. ".At
Bidim C34
20

• POLYESTER

10 t=::::::~::=~-6-a-i:J:--_-6--6-~"--~~S~t:O:bilenko T 100

TIME (minutes)
OL------.-------.------r------,-------.-------r------,---~
0.1 10 100 1000 10000 100000 1000000

FIGURE 39. Creep properties of various non-woven fabrics (Allen et al.,

1982)
 57

This figure presents results of creep tests at 50 % of breaking load,

22°C, on 152 mm wide strips. It may be seen that two of the three
polypropylene fabrics (Propex, woven, and Typar, non woven heat bonded)
were broken. The third one (Fibretex, non woven, needle punched) exhi-
bits, after 100 000 minutes (69 days), creep strains as high as 200 %.
On the contrary, polyester non woven fabrics exhibit low creep strains.
It should be noted that the needle punched Bidim C34 shows higher
instantaneous strains than the resin bonded Stabilenka T100, whereas the
creep rate of the former is much smaller than the latter's one.
In order to check the possibility of using polymers in cold regions,
these writers also studied the influence of low temperatures, and they
could observe that, whereas polyester presents few variations when
lowering the temperature from 22° to -12°C, polypropylene creep proper~
ties are radically improved. This positive influence of lowering tempe-
rature on creep behaviour of polymers is a well-known phenomenon.
Generally speaking, it may be concluded that creep properties of the
different types of geotextile used in the soil reinforcement are highly
dependent on the nature of the constituant polymer. Whereas polyester
products exhibit satisfactory behaviour, the performances of polypropy-
lene products are poor. Designers must then be careful and impose smooth
stress conditions to these reinforcements, with respect to their brea-
king strength. Further studies are then highly desirable in order to get
a precise insight on the consequences of these creep properties on the
long term stability of reinforced retaining structures.

4.3 Durability
As mentionned by many writers, one of the reasons why people started
considering polymers as a possible material for soil reinforcement was
the risks of electro-chemical corrosion encountered when using metallic
inclusions.
In a general manner, the problem of long term durability of any type
of material submitted to given physico-chemical and mechanical
conditions is very much involved, and long term real experiments seem to
be the only reliable solution. Thus, adopting a new technology in which
durability problems are involved is always dangerous, and may lead to
important errors, due to previously unknown phenomena.
For example, in the case of metallic strips, previous m~dium term
experiments developed on stressed stainless steel strips confined in-
soil gave excellent results, which led the Reinforced Earth Company to
adopt this metal for strips. Nowadays, some of the walls built with this
type of strips are presenting some problems due to a rapid and unexpec-
ted corrosion, and have to be repaired. On the contrary, galvanized soft
steel generally shows excellent behaviour, and the numerous corres-
ponding walls, which were erected 15-20 years ago, still behave
satisfactorily.
Paragraph 2.2. of this report described the use of. polymers in
Reinforced Earth Technique. In the case of the wall of Poitiers, strip
samples were taken from the upper part of the wall. According to tensile
tests, a 50 % loss of strength, and a 25 % decrease of the rupture
strain were estimated. Optical and electron micrographs evidenced an im-
portant degradation due to setting (flexion, shearing or crushing
actions). The micrographs also revealed a degradation of the strip sur-
face, corresponding to a "dirty" macroscopic aspect, which was
attributed to a contamination caused by the soil.
In fact, durability problems of polymers, and ageing phenomena are
58

increased, as in any other material, by stressing. The theme of

durability was developed in Session 8B of Las Vegas Conference (4 pa-
pers). Unfortunately no special session was dedicated to this subject at
Vienna (1986), whereas a Rilem symposium was organized in Paris (1986)
on "Long term behaviour of Geotextile". Finally, relatively few papers
deal with this important problem.
The sensitivity of polymers to Ultra Violet radiations is well known,
and Von Vijk and Stoerzer (1982) give a description of the chemical
mechanisms involved, based on the presence of oxygen, of a minimal
activation energy, and consists in an autoxydation process which finally
results in the formation of inert products. Each type of polymer is most
sensitive to a critical wave-length in the 300-370 mm range for
polyethylene, polyester and polypropylene. Intensity of radiation great-
ly influences the rate of ageing, together with temperature and humidi-
ty. Tropical conditions are for example critical. However, stabilizers
may improve resistance against day light.
Figure 40 (Raumann, 1982) shows the degradation in strength, obtained
from grab tests, which occured in a polyester non woven fabric, after a
32 weeks exposition"in Florida, Arizona and North Carolina. A considera-
ble degradation of the strength may be observed, depending on the loca-
lization of the sample.

100
GRAB STRENGTH • •+ *[;:, Florida
Arizona

o North California

60

20

WEEKS EXPOSED
8 16 24 32

FIGURE 40. Decrease of the grab strength of a non-woven fabrics sub-

mitted to light in various parts of the U.S. (Raumann, 1982)

For all these reasons, users of polymers in geotechnical engineering

are very cautious, and they never let fabrics exposed to light for long
periods. In the case of geotextile walls, protective measures, such as
bitumen projection, or non polymer facing, are adopted. Some people
consider special the design of facings, allowing periodical replace-
ments.
Other attacks on polymers may be caused by chemical and biological
action, and the effects resulting from the burial of stressed polymers
within wet soils still need to be well investigated, and require more
research.
 59

The a-priori great confidence geotextile users put in polymers may

come from the remarkable performances of PVC, since adduction of water
under pressure has been based on PVC pipes since 1935, which still
behave very satisfactorily. More recent polymers have comparable mecha-
nical properties, and considerable progress has also been made in terms
of additive agents and stabilizers. However, these good performances
should not hide some reactions polymers may have with soils.
Oxydation of polymers always has to be initiated by another agent,
such as V.V. radiation or heat, and this should not occur within a soil
mass.
Presence of water may induce some effect, corresponding to absorption
of water within the polymer structure. The intensity of this action
varies in the same way as permeability, and polyamids are known to be
the more reacting polymers to water. The absorption decreases when the
cristallinity rate increases. It may only be a physical absorption,
which induces a swelling corresponding to a finite decrease of strength.
Absorption may however by accompanied by a chemical reaction called hy-
drolisis, which modifies the structure of the polymer. This reaction is
however very slow at current temperature. Polypropylene is known to have
a good stability with respect water action.
When a polymer is buried in ground, several factors related to mine-
ral or organic action have to be considered: acid, or alcalis, soluble
salts, from iron or magnesium for instance; organic elements, such as
acids which come from bacteriologic destruction of organic matter.
Generally, except for polyamids, current polymers have a satisfactory
resistance to those agents. Polyester may be more sensitive to alkalis.
Biological action, as coming either from bacteria or fungus may
affect polymers. This action is better known for drainage problems,
since it corresponds to a biocolmatation caused by microbial produced
ferric hydroxide, ferrous sulfide and polysaccharide. However, as
mentionned previously in the case of the Reinforced Earth wall built
with fiberglass reinforced plastic strips, biological action may be
quite agressive, fast and unpredictable. Ionescu et al. (1982) have
performed interesting tests by incubating various woven and non woven
polyester and polypropylene fabrics in eight different media, including
distilled water, iron bacteria, desulfovibrios, levan synthesiying bac-
teria, sea water, mineral solution, compost and alluvial soil. After 5
and 17 months, they observed some bio-colmatation, but an i~significant
change in the tensile strength values. It should be noted however that
no stress were applied during those experiments, and that very few
results on the chemical and biological behaviour of stressed polymers
buried in such media exists. Furthermore, as noted by Mallinder et al.
(1977), the 6 months duration laboratory tests performed on fiberglass
plastic strips were also satisfactory. Extrapolation of those kind of
tests for long term behaviour should then be made very cautiously.
It is generally believed that long term stressing of polymers may
induce some superficial microfissurations which may considerably favo-
rize chemical or bio-chemical attacks, through the microdefects de-
veloping at the surface. This development of microfissures is much more
important for tensile stresses than for compression stresses.
The best way of understanding those problems is to perform full scale
experiments, which consist for instance in observing polymers after a
long period of true working conditions. An important work, described by
Sot ton et al. (1982), has been undertaken in this direction by the
French Committee of Geotextiles. This study concerns different types of
60

woven and unwoven fabrics, made of polypropylene and polyester fila-

ments. Many observations have been made, and mechanical control tests
were performed. Writers insist on the need of conserving reference
samples of the geotextiles used, in order to make valid comparisons,
particularly in terms of the evolution of the mechanical properties.
They observe few variation when the geotextile is not submitted to
stress. Exposure to weather variations and geotextile tearing induced
large decreases in strength ( > 30 %). This may happen when the geotex-
tile stays too long without being covered by a soil layer. For the
application to retaining structures, it seems interesting to rather
consider, with respect to creep consequences, the studied cases where
the geotextiles were not apparently damaged, but had to support a loa-
ding. In this perspective, it can be refered to the case histories of
Noyalo (non woven needle punched polypropylene) and Thiers (woven
polypropylene). , The Noyalo fabric appeared to be contaminated by fine
grained soils, and suffered a 20-40 % loss of strength. This contami-
nation could be compared to the degradation of the strips of the
Poi tiers wall. The case of Thiers, which is the only case where the
fabric had to sustain a 5 m high embankment, seems more serious, with a
53 % loss strength, attributed by the authors to creep under the
considered load.
Such long term applications of load seem to be an important factor of
strength reduction, because of creep. This aspect is particularly impor-
tant in the case of retaining structures, where fabrics are submitted to
important loads. For this reason, the problem of long term stability of
these structures still has to be discussed regarding polymer properties.
In this perspective, the study of the properties of the Bidim fabric of
the first geotextile reinforced wall of Rouen (1971) will be extremely
interesting. Such a study has been started by the Laboratoire des Ponts
et Ghaussees (L.P.G.) and the general aspect of the fabric, after 16
years in the ground, is good (Helmas and Gourc, 1987). Mechanical
data have now to be precisely studieu in order to draw valuable con-
clusions. On the other hand, the behaviour of the various fabric rein-
forced soil walls monitored by LPG in France for some years seem at
present satisfactory, in terms of external deformations as a function of
time.
However, predictions are difficult to make. In fact, the only
suitable method to answer the difficult problem of 'durability is the
full scale long term experimentation. Before drawing any valuable con-
clusions from this kind of approach, relative caution in any assessment
concerning long term performance of any material as a reinforcement
material seems to be desirable.

GENERAL CONCLUSIONS
Polymers are or can be used in numerous types of retaining systems,
ranging from the classical "ladder wall" invented by Coyne to Texsol,
recently developed by Leflaive. Of course, each application is different
and depends on many factors (cost, aesthetic aspect of the structure,
structure deformability, durability, etc ... ).
' As used in reinforcements, polymers provide an alternative to metals,
the latter being so far limited to linear inclusions and grids. However,
their deformability is generally higher than metals and it appears to be
a disavantage. They are not subject to corrosion, but they can be degra-
ded.
Considering wall construction, the use of deformable polymers
 61

generally requires special types of facing in order to built vertical

walls.
For reinforced soil walls, in which the frictional soil-inclusion in-
teraction is predominant, the use of polymers has permitted to cover all
the range of reinforcement types 1) linear reinforcement with
materials like fiber reinforced plastic; 2) two-dimensional reinforce-
ment with geotextiles and geogrids; 3) three-dimensional reinforcement
with a continuous thread (Texsol), fibers or mesh elements, but also
with combinations of grids like for gabions, rafts, or mattresses.
Future development in the use of polymers for retaining walls cons-
truction requires further research in the following areas
- Reinforced soil mechanical behaviour and durability.
- Design methods taking into account deformations.
- Facing construction techniques.

REFERENCES

1. Bacot J: Contribution a l'etude du frottement entre un materiau

souple et un materiau pulverulent. These de Doctorat es Sciences. INSA
de Lyon, 1981.
2. Baudonnel J, Giroud JP, Gourc JP: Etude experimentale et theorique
du comportement en traction des geotextiles non tisses, Proceedings of
the 2nd International Conference on Geotextiles, volume 3, pp. 823-828,
Las Vegas, 1982.
3. Bassett RH, Last HC: Reinforced Earth below Footings and
Embankments. ASCE Symposium on Earth Reinforcement, pp. 222-231,
Pittsburgh, April 1978.
4. Bell AL, Green HM, Laverty K: Factors Influencing the Selection of
Woven Polypropylene Geotextile for Earth Reinforcement. Proceedings of
the 2nd International Conference on Geotextiles, volume 3, pp. 689-694,
Las Vegas, 1982.
5. Blivet J.C., Grestin F.: Etude de l'Adherence entre Ie Phosphogypse
et deux Geotextiles. ColI. Int. Renf. des Sols, Vol. II, Paris, pp.
403-408, 1979.
6. Bonaparte R, Kamel MI, Dixon JH: Use of geotextiles in soil reinfor-
cement. Proceedings of the 1984 annual meeting of the Transportation
Research Board, Washington DC, January 1984.
7. Bonazzi D., Colombet G.: Reajustement et Entretien des A~crages de
Talus. Proc. of the Int. Conf. on In-Situ Soil and Rock Reinforcement,
Ecole Nationale des Ponts et Chaussees, Paris 9-11 Octobre 1984, pp.
225-230, 1984.
8. Cazzuffi D, Venezia S, Rinaldi H, Zocca A: The Mechanical Properties
of Geotextiles Italian Standard and Interlaboratory Test Comparison.
Proceedings of the 3rd International Conference on Geotextiles, volume
3, pp. 879-884, Vienna, 1986.
9. Chabard J.P., Pardieu P., Guerber P., Bertrand J.: Novel Reinforced
Fill Dam. Proc. 8th ECSMFE, Elsinki, Session 5, 1983.
10. Chang JC, Hannon JB, Forsyth RA: Pull Resistance and Interaction of
Earthwork Reinforcement and Soil. Proceedings of the 1977 Meeting of
the Transportation Research Board. Washington DC, January 1977.
11. Coyne A.: Murs de Soutenement et Murs de Quai "a Echelle", Le Genie
Civil, 1er et 15 Mai, 1945.
12. Delmas P, Berche JC, Gourc JP: Le dimensionnement des ouvrages
renforces par geotextiles: Programme CARTAGE. Bulletin de Liaison des
Ponts et Chaussees, nO 142, pp. 33-44, Paris, Mars-Avril 1986.
62

13. Delmas P, Gourc JP: Le Renforcement par Geotextiles: Recherches et

Realisations. Reunion du Comite Fran~ais de Mecanique des Sols, Paris,
Avril 1987.
14. Deimas P, Puig J, Schaeffner M: Mise en oeuvre et parement des
massifs de soutenement renforces par des nappes de geotextiles. Bulle-
tin de Liaison des Ponts et Chaussees, nO 143, pp. 65-77, Paris, Mai-
Juin 1986.
15. Finnigan JA: The Creep Behaviour of High Tenacity Yarns and Fabrics
Used in Civil Engineering. Proceedings of the International Conference
on the Use of Fabrics in Geotechnics, volume 2, pp. 305-310, Paris,
1977 .
16. Forsyth RA: Alternative Earth Reinforcements. Proceedings of the
ASCE Symposium" on Earth Reinforcement, pp. 358-370, Pittsburgh, April
1978.
17. Forsyth RA, Bieber DA: La Honda Slope Repair with Geogrid Reinfor-
cement. Proceeding~ of the Symposium on Polymer Grid Reinforcement, pp.
54-57, London, 1984.
18. Fukuoka~, Imamora Y: Fabric retaining walls. Proceedings of the
2nd International Conference on Geotextiles, Volume 3, pp. 575-580,
Las Vegas, 1982.
19. Gasnier R, Plumelle C: Etude experimentale en vraie grandeur de
tirants d'ancrage. International Conference on in-situ Soil and Rock
Reinforcement, Ecole Nationale des Ponts et Chaussees, pp. 333-339,
Paris, 1984.
20. Giroud JP, Carroll Jr: Geotextile Products. Geotechnical Fabrics
Report, volume 1, n O l, IFAI, pp.12-15, St Paul, Minnesota, 1983.
21. Gray DH: Role of woody vegetation in reinforcing soils and stabili-
zing slopes, Symposium on soil reinforcement and stabilizing
techniques, pp. 253-306, Sydney, ·1978.
22. Gray DH, Athanasopoulos G, Ohashi H: Internal-External Fabric
Reinforcement of Sand. Proceedings of the 2nd International Conference
of Geotextiles, volume 3, pp. 611-616', Las Vegas, 1982.
23. Gray DH, Ohashi H: Mechanics of fiber reinforcement in sand, ASCE,
Journal of the Geotechnical Engineering Division, (109) GT3, pp. 335-
353, 1983.
24. Greenwood JH, Myles B: Creep and Stress Relaxation of Geotextiles.
Proceedings of the 3rd International Conference on Geotextiles, volume
3, pp. 821-826, Vienna, 1986.
25. Hausmann MR: Strength of Reinforced Soil. Australian Road Research
Board, Proceedings, (8), 1-8, Session 13, 1976.
26. Hoare DJ: Laboratory study of granular soils reinforced with
randomly oriented discrete fibers. Proceedings of the International
Conference on Soil Reinforcement (1), pp. 47-52, Paris, 1979.
27. Ingold TS: A laboratory investigation of grid reinforcements in
clay. Geotechnical Testing Journal, GTODJ, volume 6, n° 3, September
1983.
28. Ionescu A, Kiss S, Dragan-Bularda M, Rodulescu D, Kolozsi E, Pintea
H, Crisan R: Methods used for Testing the Bio-Colmatation and Degrada-
tion of Geotextiles Manufactured in Romania. Proceedings of the 2nd
International Conference on Geotextiles, volume 3, pp. 547-552, Las
Vegas, 1982.
29. Jewell RA: Some effects of reinforcement on the mechanical beha-
viour of soil. PhD thesis, University of Cambridge, 1980.
30. Jewell RA, Jones CJFP: Reinforcement of clay soils and waste mate-
rials using grids. Proceedings of the 10th International Conference of
 63

Soil Mechanic and Foundation Engineering (3), pp. 701-706, Stockholm,

1981.
31. John N.W.H., Ritson R., Johnson P.B., PETLEY D.J.: Instrumentation
of Reinforced Soil Walls. Proceedings of the 8th ECSMFE. Helsinki , May
1983, Vol. 2, pp. 509-512, 1983.
32 . Jones CJFP: Practical design considerations. Proceedings of the
Symposium on Reinforced Earth and Other Composite Soil Techniques. TRRL
and Heriot-Watt University, September 1977.
33. Jones CJFP: Practical Construction Techniques for Retaining
Structures using Fabric and Geogrids. Proceedings of the 2nd Interna-
tional Conference on Geotextiles, volume 3, pp. 581-585, Las Vegas,
1982.
34. Kern F: Realisation d'un barrage en terre avec parement aval verti~
cal au moyen de poches en textile. International Conference on the Use
of Fabrics in Geotechnics. Ecole Nationale des Ponts et Chaussees,
volume 1, pp. 91-94, Paris, 1977.
35. Leclercq B et Prudon R: Comportement en traction des geotextiles en
fonction de l'inclinaison relative de l'axe de production par rqpport a
l'axe de l'effort, Proceedings of the 3rd International Conference on
Geotextiles, volume 3, pp. 751-756, Vienna, 1986.
36. Leflaive E, Paute JL et Segouin M: La Mesure des Caracteristiques
de Traction en vue des Applications Pratiques, Proceedings of the 2nd
International Conference on Geotextiles, volume 3, pp. 733-738, Las
Vegas, 1982.
37. Leflaive E, Khay M, Blivet JC: Un nouveau materiau : le Texsol.
Bulletin de Liaison du Laboratoire des Ponts et Chaussees, nO 125, pp.
105-114, Paris, Mai- Juin 1983.
38. Leflaive E, Liausu PH: Le Renforcement des Sols par Fils Continus.
Proceedings of the 3rd International Conference on Geotextiles, volume
2, pp. 523-529, Vienna, 1986.
39. Mallinder F.P.: The Use of FRP as Reinforcing Elements in Reinfor-
ced Soil Systems. Proceedings of the Symposium on Reinforced Earth and
other Composite Soil Technique. TRRL and Heriot-Watt University,
Edinburgh, 1977.
40. McGown A, Andrawes KZ, Al Hasani MM: Effect of Inclusion Proper-
ties on the Behaviour of Sands. Geotechnique (28), 3, pp. 327-346,
1978.
41. McGown A, Andrawes KZ, Kabir MH: Load-extension Testing ' of Geotex-
tiles Confined in-soil, Proceedings of the 2nd International Conference
on Geotextiles, volume 3, pp. 793-798, Las Vegas, 1982.
42. McGown A, Andrawes KZ, Yeo KC: The Load-Strain-Time Behaviour of
Tensar Geogrids. Proceedings of the Conference on Polymer Grid Reinfor-
cement, pp. 11-17, London, 1984.
43. Mercer FD, Andrawes KZ, McGown A, Hytiris N: A new method of soil
stabilization. Proceedings Symposium on Polymer Grid Reinforcement in
Civil Engineering. Paper 8-1, London, 22-23, March 1984.
44. Morbois A., Long N.T.: Etude du procede Actimur. Rapport de Recher-
che. Laboratoire Central des Ponts et Chaussees, Paris, 1984.
45. Murray RT, Irwin MJ. A Preliminary Study of TRRL Anchored Earth.
TRRL supplementary report 674, 1981.
46. Murray RT, McGown A, Andrawes KZ, Swan D: Testing Joints in Geotex-
tiles and Geogrids, 3rd International Conference on Geotextiles, volume
3, pp. 731-736, Vienna, 1986.
47. Paute JL, Segouin M: Determination des caracteristiques de
resistance et de deformabilite des textiles par dilatation d'un manchon
64

cyclindrique, International Conference on the Use of Fabrics in

Geotechnics, volume 4, pp. 293-298, Paris 1977.
48. Puig J, Blivet JC, Pasquet P: Remblai arme avec un textile synthe-
tique. International Conference on the Use of Fabrics in Geotechnics.
Ecole Nationale des Ponts et Chaussees, volume 1, pp. 85-9, Paris,
1977 .
49. Rabejac S, Toudic P: Construction d'un mur de soutenement entre
Versailles-Chantiers et Versailles-Matelots. Revue Generale des Chemins
de Fer, 93eme annee, 1975.
50. Raumann C: A Hydraulic Tensile Test with Zero Transverse Strain for
Geotechnical Fabrics. Geotechnical Testing Journal, volume 2, n03, pp.
69-76, June 1979.
51. Richards EA, Scott JD: Stress-strain Properties of Geotextiles.
Proceedings of the 3rd International Conference on Geotextiles, volume
1, pp. 873-878, Vienna, 1986.
52. Rowe RK, Ho SK: Determination of Geotextile Stress-strain Characte-
ristics Using a Wide Strip Test. 3rd International Conference on
Geotextiles, volume 3, pp. 885-890, Vienna, 1986.
53. Schlosser F: Discussion Session. International Conference on the
Use of Fabrics in Geotechnics, volume 3, pp. 36-37, Paris, 1977.
54. Schlosser F, Long NT: Comportement de la Terre Armee dans les
Ouvrages de Soutenement. Proceedings of the 5th European Conference on
Soil Mechanics .and Foundation Engineering, volume 1, pp. 299-306,
Madrid, 1972.
55. Schlos,ser F, Elias V: Friction in Reinforced Earth. Proceedings of
the ASCE Symposium on Earth Reinforcement, pp. 735-764, Pittsburgh,
April 1978.
56. Schlosser F, Jacobsen HM, Juran I: Soil 'Reinforcement. General
Report. Speciality Session 5, Proceedings of the 8th European Conferen-
ce on Soil Mechanics and Foundation Engineering (3), pp. 1159-1180,
Helsinki, 1983.
57. Schlosser F, Magnan JP, Holtz RD: Geotechnical Engineered
Construction. Theme Lecture 5, Proceedings of the 11th International
Conference on Soil Mechanics and Foundation Engineering (1), pp. 211-
254, San Francisco, 1985.
58. Schwantes ED, JR: Recent Experience with Fabric-faced Retaining
Walls, Proceedings of the 2nd International Conference on Geotextiles,
volume 3, pp. 605-609, 1982.
59. Shresta SC, Bell JR: A Wide Strip Tensile Test of Geotextiles,
Proceedings of the 2nd International Conference on Geotextiles, volume
3, pp. 739-744, Las Vegas, 1982.
60. Sissons CR: Strength testing of fabrics for use in civil enginee-
ring. International Conference on the Use of fabrics in Geotechnics,
volume 2, pp. 287-292, Paris, 1977.
61. Sotton M, Leclercq B, Paute JL, Fayoux D: Quelques Elements de
Reponse au Probleme de la Durabilite des Geotextiles. Proceedings of
the 2nd International Conference on Geotextiles, volume 3, pp. 553-558,
Las Vegas, 1982.
62. Van Leeuwen JH: New methods of determining the Stress-strain Beha-
viour of Woven and Non Woven Fabrics in the Laboratory and in Practice.
International Conference on the Use of Fabrics in Geotechnics, volume
IV, pp. 299-304, Paris, 1977.
63. Van Vijk W, Stoerzer M: UV Stability of Polypropylene. Proceedings
of the 3rd International Conference on Geotextiles, volume 3, pp. 851-
856, Vienna, 1986.
 65

64. Vautrain J, Puig J: Experimentation du Bidim ; Remblai Experimental

de Caen. Bulletin de Liaison des Laboratoires Routiers des Ponts et
Chaussees, nO 41, Paris, Novembre 1969.
65. Verma BP, Char ANR: Triaxial Tests on Reinforced Sand. Proceedings
of the Symposium on Soil Reinforcing and Stabilizing Techniques, pp.
29-39, Sydney, 1978.
66. Vidal H: La terre armee (un nouveau materiau pour les travaux
publics). Annales de l'ITBTP, nO 223-224, Juillet-Aout 1966.
67. Viergever MA, De Feijter JW, Mouw KAG: Biaxial Tensile Strength
and Resistance to Cone Penetration of Membranes. International
Conference on the Use of Fabrics in Geotechnics, volume II, pp. 311-
316, Paris, 1977.

ACKNOWLEDGEl1ENTS
The authors acknowledge Mr J. CANOU for his help in preparing the ·En-
glish version of the text.
Prediction Exercise
INTRODUCTION AND RATIONALE FOR PREDICTION EXERCISE

PETER M. JARRETT

Royal Military College of Canada

Two large scale polymerically reinforced soil walls were built,

instrumented and tested at the Royal Military College of Canada. The
des ign of the experiments, which represented working levels of stress,
arose as a result of discussions held under a NATO Collaborative Re'search
Grant between Drs. McGown and Andrawes of the University of Strathclyde in
Scotland and Drs. Jarrett and Bathurst at the Royal Military College of
Canada.

Prior to construction, details of the proposed tests and the materi-

als to be used were sent to potential predictors. They were asked to make
Class A (in advance of construction) predictions of aspects of the test
behaviour such as the force on and the displacements of the facing
elements, the strains in the reinforcement, the earth pressure distribution
and a projected failure load. In addition they were asked to provide
details of their analytical methods. The understanding from the outset of
the exercise was that the predictors would meet to discuss in a friendly,
cooperative manner the measured results and the predictions. It was also
stated that to allow freedom to experiment analytically and to maintain
openess of discussion that there would be no formal attribution of the
predictions to their predictors. The only competitive aspect was the prize
of a bottle of Teachers Whiskey provided for the closest prediction
courtesy of our Scots colleagues.

The following persons submitted predictions. The list is in Alpha-

betic order only and bears no relationship to the order of discussion of
predictions in a later paper:

Rudolph Bonaparte
Danielle Cazzuffi, Pietro Rimoldi and Roberto Mangiavacchi
Philippe Delmas and Jean-Pierre Gourc
Richard Jewell
Colin Jones
Robert Koerner and Manfred Hausmann
Richard Murray
Gregory Richardson
Jost Studer and B. Graf

The organizers most sincerely thank these gentlemen for their consi-
derable efforts, their spirit in accepting the challenge and for the
friendly, open and cooperative manner in which their discussions were
carried out.

The outcome of the prediction exercise is presented in the following

chapter together with certain amplifying material that it is felt increases

69

P. M. Jarrett and A. McGown (eds.), The Application ofPolymeric Reinforcement in Soil Retaining Structures, 69-70.
© 1988 by Kluwer Academic Publishers.
70

the overall value of the basic test program. There are two keystone papers
for the prediction exercise and three supplementary papers.

Comprehensive details of the construction of the RMC Test Walls and

the measured results are provided in the first paper by Bathurst, Wawrychuk
and Jarrett. Most of the information provided to predictors is contained
in this paper to enable readers to test their own analytical methods
against the results. The second paper by Bathurst and Koerner presents
details of the methods employed by predictors and a comparison of the
results of the predictions to the measured behaviour. It makes fascinating
reading!

The three papers that follow contain most valuable supplementary

information and analyses that were completed after the workshop.

During demolition of the test walls, samples of the Tensar SR2 rein-
forcement were recovered and returned to the manufacturer, Netlon Ltd. for
examination and testing to assess the material's resistance to damage
during construction. The third paper by Bush and Swan reports the results
of this work.

It was suggested during the workshop that measurements made on a

propped wall, retaining unreinforced soil might help in the interpretation
and comprehension of. the reinforced soil results. Most especially it could
shed light on the effect of side wall friction in the test facility and the
actual angle of friction of the soil. A preliminary test of this nature
was carried out after the workshop together with further measurements of
the side wall friction potential. These results are presented by Bathurst
and Benj amino

The final paper by Jewell represents a very thorough analysis of the

Test Walls. It is especially valuable in the context of this chapter for
its interpretation of the strength properties of the sand used in the Test
Walls and for the analysis of the effects of sidewall friction on the earth
pressures in the RMC Test Facility. The paper also acts as a test for and
example of the use of the analytical procedures proposed by Jewell in a
companion paper in this volume.

In summary, it is believed that this chapter contains a detailed

case history of two reinforced soil walls that may be used as a benchmark
test for analytical methods as they are developed and it provides a graphic
state of the art of the accuracy of the present analytical methods used by
the predictors. There is still much to be learned about both soil and
polymer behaviour.
LABORATORY INVESTIGATION OF TWO LARGE-SCALE
GEOGRID REINFORCED SOIL WALLS

Richard J. Bathurst, William F. Wawrychuk and Peter M. Jarrett

Civil Engineering Department

Royal Military College of Canada

1.0 Introduction

1.1 General
Two(2) large-scale model reinforced soil walls were constructed within the soil
retaining wall test facility at the Royal Military College, Kingston, Canada.
The vertical reinforced soil walls were 3 m high and were constructed using:
a) an incremental timber panel facing with a polymer geogrid soil reinforcement, and;
b) a propped (tilt-up) timber panel facing with a polymer geogrid soil reinforcement.
Each wall was subjected to sustained surcharge loadings up to 50 kPa following
construction.

1.2 Objectives
The construction and performance monitoring of the two reinforced soil struc-
tures was undertaken to provide a focal point for discussion at the NATO Advanced
Research Workshop entitled: Application of Polymeric Reinforcement in Soil Retaining
Structures held at the Royal Military College of Canada in June 1987.
The workshop was broadly directed to review the state-of-the-art in polymer rein-
forced soil retaining structures and to critically assess the relationships between design
methodologies, construction practices and the measured performance of these aystems.
In order to stimulate discussion, invited experts were given details of the proposed
reinforced soil wall configurations before construction (Bathurst and Jarrett, 1986)
and asked to make Glass A predictions of the behaviour of the two reinforced soil
structures at various elapsed times after initial construction and after surcharging
pressures were applied to the structures. The results of the prediction exercise were
presented informally and were the centre of discussion for two workshop sessions.
This paper describes the construction of the two test walls and the measured test
results. A summary of the performance predictions offered by Workshop participants
is reported by Bathurst and Koerner in these proceedings.

2.0 RMC Retaining Wall Test Facility

2.1 General
The RMC Retaining Wall Test Facility was conceived and constructed to provide
a general purpose large-scale test apparatus to examine a variety of reinforced soil wall
systems. The test facility is located within the Dolphin Structures Laboratory of the
Civil Engineering Department at RMC. The principal structural components of the
test facility were completed in November 1985.

71

P. M. Jarrett andA. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 71-125.
© 1988 by Kluwer Academic Publishers.
72

2.2 Description of Test Facility

An overview of the RMC Retaining Wall Test Facility is given on Plate 1. The
principal structural components of the facility are six rigid heavily reinforced concrete
counterfort cantilever wall modules which are shown on Figure 1. These modules are
arranged to laterally confine a block of soil up to 6.0 m long by 3.6 m high by about
2:.4 m wide. Translational stability of the reinforced concrete segments is provided by
32 mID dia. anchor rods/bolts extending through bolt holes to structural anchors in
the laboratory floor. Additional lateral stability is provided by six rectangular hollow
structural sections bolted across the top of wall segments at the location of the module
counterforts.
The back of the soil block between concrete wall segments is confined by a system
of braced vertical posts and lagging boards. Reinforced soil wall facings are exposed
at the front edge of the test facility.
Surcharging of test configurations is carried out by inflating reinforced air bags
confined between the concrete walls and a timber ceiling. The timber ceiling is re-
strained in turn by the hollow structural steel sections as shown on Figure 2. The
current surcharging arrangement allows a vertical pressure equivalent to 3 m of fill to
be applied to the soil surface between the concrete modules. The inside walls of the
structure are faced with a composite plywood/clear plexiglas/polyethylene sheeting
which acts to reduce sidewall friction. In addition, the plywood has been drilled out at
regular spacings to allow the soil fill and reinforcing elements to be observed directly
during and after construction of test models.
Additional details of the test facility have been reported by Lescoutre (1986) and
Wawrychuk (1987).

3.0 Test Configurations

3.1 General
Schematics of the reinforced soil walls tested are shown on Figures 3 and 4. Both
constructions comprised a high density polyethylene TENSAR SR2 Geogrid-reinforced
sand soil. Reinforcement layers were attached to the plywood bulkhead facing panels
shown on the figures. The principal differences between the structures were the front
panel configurations and details of construction sequence. The incremental facing panel
wall was constructed in four stages with each panel externally supported only until the
sand fill behind the panel was placed and compacted. The facing panel supports on
the propped wall construction, however, were released only after the full height of
sand fill had been placed behind the facing units. Both model Vl{alls were constructed
with a central instrumented section nominally 1 m wide and two 0.7 m wide edge
sections. This construction was adopted to reduce edge-effects on the performance of
the monitored central reinforcement strip and facing unit(s). A foam rubber void filler
was placed along all panel edges in order to prevent the panels from binding during
outward movements. In addition, for the incremental wall test, a layer of compressible
foam rubber was placed at each horizontal panel to panel interface to assist in panel
levelling during construction of the wall.
A principal objective of the current investigation was to examine the behaviour
of these reinforced soil wall systems under working stress conditions. Therefore, the
dimensions of the trial sections were based essentially on the design methods contained
in the U.K. Department of Transport's technical memorandum Reinforced Earth Re-
taining Walls and Bridge Abutments for Embankments, BE 9/78 with one exception,
the grid lengths were limited to 3 metres instead of 5 metres. Preliminary trial tests at
RMC with wrap-around geogrid facings have shown that for the same number of rein-
forcing layers (e.g. four) and a total height offill equal to 3.25 m, 3 metre reinforcement
lengths are more than adequate for stability (Bathurst, 1987).
 73

Plate 1 Overview of RMC Retaining Wall Test Facility

74

STRUCTURAL

0
4m

0 0 0 MODULE
NUMBER
0 0 0 0

0
0 0 0 0
0 0
0
0 0 0 0 0
0

0 0 0 0

ANCHOR BOLTS

Figure 1 RMC Retaining Wall Test Facility

FRONT OF WALL
HSS 152x102x9.5
(
TIMBER BEAM
38 x 184 (TYPJ

t------~*-
SURCHARGE
PANEL

SETTLEMENT
.PLATE ________ 1 ___ _
SAND SURCHARGE 200

REINFORCED SOIL
FRONT PANEL

ALL DIMENSIONS IN mm.

Figure 2 Air Bag Surcharging System

 75

SAND/AIR PRESSURE BAG SURCHARGE

1 24m
--j
SURCHARGE
PANELS
n
/1 \ -

T E
SR2 LAYER 3

I
SR2 LAYER
2 ""j:af'"'----~--"........"""-t E

SR2 LAYER I
~L~~N~ ~y~ ~ _ _ ... _______ ....
~
l
• STRAIN GAUGE
ErE LEVELLING PAD"
_ PRESSURE CELL

CROSS - SECTION VIEW FRONT VIEW

Figure 3 General Arrangement for Incremental Facing Wall Construction

SAND/AIR PRESSURE BAG SURCHARGE r- 2.4m--j

D\ ,
SURCHARGE
PANELS

, 1\
T
SR2 LAYER 3
~~
SR2 LAYER 2 ~
~
"
z
u
1t

~
l
ETE LEVELLING PAD ..:.
• STRAIN GAUGE
_ PRESSURE CELL

CROSS -SECTION VIEW FRONT VIEW

Figure 4 General Arrangement for Propped Facing Wall Construction

Details of wall design and construction were arrived at through collaboration be-
tween the authors and Dr. A. McGown, Dr. K. Andrawes and D. Varney at the
University of Strathclyde, Scotland. Construction details are considered realistic for
actual field structures although some compromises in construction were made to make
the performance· of these structures easier to predict and monitor. For example, facing
panels are not dovetailed or staggered and the facings were arranged without an initial
batter in the vertical direction. Another example is the panel/reinforcement connec-
tions. The number of bolts per grid was limited to four in order to ensure adequate
sensitivity of the bolts as load cells.
76

3.2 Materials
3.2.1 Sand
The soil used was a washed sand with some gravel. This material is composed of
sub-angular to angular quartz and feldspar particles. The grain size distribution for
this material is given on Figure 5. Less than 0.3% of all particles by weight are less
than 0.075 microns.
The results of direct shear tests and triaxial compression tests carried out at RMC
are given in Appendix A. A summary of test results is given on Table 1. The second
column of friction angles shown on the table are values representing linear failure
envelopes passing through the origin of a T - Un plot. Additional direct shear tests on
the RMC sand are reported by Jewell (1987b) as part of these proceedings and confirm
(secant) peak friction angles interpreted from the RMC direct shear tests.

SIEVE SIZE IN mm.

0.1
100

90

80

(!) 70
Z
en
en 60
«
a..
~ 50
Z
I.&J
~ 40
I.&J
a..
30

20

10

0
200
U.S. SIEVE Nil.

Figure 5 Sand Fill Grain-Size Distribution

3.2.2 Sidewall Friction

A preliminary, large-scale shear box test (1 m by 1 m in plan view) gave a fully
mobilized friction angle of 20° for the sand/sidewall interface. The results of a more
recent and comprehensive set of direct shear box tests reported in this proceedings
by Bathurst and Benjamin shows that 20° may be an upper-bound value. A value
of 15° is likely a better estimate of the sidewall friction angle at normal stress levels
considered to act against the test facility sidewalls.
 77

Table 1

Results of Shear Strength Tests on Sand Backfill

Type of Test Dry Soil Density Friction Angle

Pd (kg/m 3 ) (degrees)
Direct Shear Box 1690 42* 44**
(6cm x 6cm x 3.7cm high)
Direct Shear Box 1780 43* 48**
(6cm x 6cm x 3.7cm high)
Direct Shear Box 1860 (peak) 46* 53**
(6cm x 6cm x 3.7cm high) (constant volume) 40* 44**
Consolidated-Drained 1680 '41 **
(Standard) Triaxial Test

* as provided to predictors in Bulletin 1, (Bathurst and Jarrett, 1986)

** based on a linear failure envelope with c=O

3.2.3 Geogrid Reinforcement

Four layers of reinforcement comprising 3 m long strips of high density polyethy-

lene TENSAR SR2 Geogrid were used in test configurations. The mechanical proper-
ties of samples taken from the actual rolls of reinforcement were determined from tests
carried out by NetIon Ltd. prior to wall construction. These results are presented in
Appendix B.

4.0 Construction Methods and Surcharge Loading

4.1 General

The construction sequence adopted for both walls was similar with the exception
of the temporary waling support and release sequence.
Prior to construction of both test configurations, a 250 mm thick blinding layer
was placed and compacted behind the concrete floor levelling pad shown on Figures 3
and 4. The timber facing panels were seated on the concrete levelling pad and initially
restrained between the concrete lip shown on the figures and a wooden waling located
in front of the wall at the same elevation. Subsequently, all sand backfill was placed and
compacted in 125 mm lifts covering the full length of the test facility. A vibrating plate
tamper was the principal means of compaction although a hand-held Kango hammer
with tamper attachment was employed to compact sand in the corners formed by the
facing panE,lls and the test facility walls.
Prior to placing fill over any strip of reinforcement, a light pretensioning load of
about 0.4 kN /m was applied to the geogrid to ensure that the reinforcement was free
of warps. The pretensioning was applied using a bar threaded through the free end of
each strip of reinforcement and attached in turn by a cable to a system of weights at
the back of the test facility. The pretensioning was released after the grid was covered
by 0.5 m of compacted sand.
78

4.2 Incremental Panel Facing Reinforced Soil Wall Test

In this construction a total of twelve 0.75 m high timber bulkhead facing panels
were used to construct a reinforced soil wall 3.0 m high. Temporary lateral support
was provided to each row of panels by a pair of horizontal timber walings braced in
turn by a system of wedging plates bolted to the front of the RMC test facility. Each
panel row support was released after the sand backfill had been placed and compacted
to the elevation of the top of the facing panel. The purpose of this form of construction
was to progressively mobilize the inherent tensile capacity of the geogrid reinforcement
as the height of the composite structure was increased.
The construction sequence is illustrated on Figure 6. Also identified on the figure
are critical construction events. A frontal view showing the incremental panel wall
facing units at the end of construction is given on Plate 2. Following construction,
the incremental facing wall was subjected to a series of surcharge loadings using a
sand surcharge or sand surcharge/pressurized airbag combination. The surcharging
schedule is given in Table 2. The design surcharge of 12 kPa was sustained for a period
in excess of 1000 hrs in order to observe creep behaviour under working (i.e design)
loads. Later, a maximum surcharge load of 50 kPa was applied for 500 hrs to observe
further creep deformations in the composite system.

Table 2

Surcharging Schedule for Incremental Panel Wall

Event Elapsed Duration (hrs)

time (hrs)
wall construction commences 0 225
12 kPa surcharge applied 225 1668
3.3 kPa surcharge during 1893 938
airbag construction
12 kPa surcharge applied 2831 24
30 kPa surcharge applied
40 kPa surcharge applied
2855
2978 . 123
101
50 kPa surcharge applied 3079 500
30 kPa surcharge applied 3579 22
12 kPa surcharge applied 3601 22
3.3 kPa surcharge applied 3623 19
excavation commences 3642 76
excavation complete 3718 -

4.3 Propped Panel Facing Reinforced Soil Wall Test

This construction comprised three(3) 3.0 m high timber bulkhead facing panels.
The facing panels were temporarily supported at the wall base, and at 1.0 m, 1.75
m and 2.50 m above the base of the wall by the same system of horizontal timber
walings described earlier. The walings were released simultaneously only after the
sand backfill had been placed and compacted to the full panel height of 3 m. The
 79

Construction Events

(!) 250 mm blinding layer placed and compacted

® Geogrid LAYER 1 in place 10
@
o
P anel 1 bracing removed
4.0
Geogrid LAYER 2 in place
® Panel 2 bracing removed
@
o
3.5 Geogrid LAYER 3 in place
Panel 3 bracing removed
@ Geogrid LAYER 4 in place
3.0
8
@ Panel 4 bracing removed

..CJ
cd
@ 750 mm uncompacted sand surcharge in place
"00 2 .5
Q)

~
..CJ
Q)
>
0
..CJ 2.0
cd
,::
0
.~
>
Q) 1.5
OJ
0;:1

1.0

I
100 150 200 250
time from start of construction (hrs)

Figure 6 Incremental Panel Wall Construction Sequence

construction sequence for this wall including critical construction events is given on
Figure 7. In a similar manner to the incremental panel wall, the propped wall was
surcharged using the airbag system installed at the top of the RMC test facility. The
surcharging schedule is given in Table 3. In this test the maximum surcharge pressure
of 50 kPa was sustained for 1000 hrs.

4.4 Sand Fill Compaction Results

Figures 8a and 8b show sand fill density and moisture content profiles as measured
during construction and at wall excavation using a Troxler nuclear density meter.
The average bulk density for the two walls was typically 1.79Mg/m3 (dry density
1.74Mg/m3 ) at the end of construction. The sand backfill was placed at a relative dry
density of about 40% and a moisture content of about 3%. The figures also show that
some densification of the soil occurred as a result of soil self-weight and surcharging.
As a result, the relative dry density is about 50% at the end of the tests.
80

Plat'e 2 Front View of Incremental Panel Wall Test

 81

Construction Events

3.5 CD 250 mm blinding layer placed and compacted

® Geogrid LAYER 1 in place
® Geogrid LAYER 2 in place
3.0
@ Geogrid LAYER 3 in place
@ Geogrid LAYER 4 in place
2.5 ® wedges removed from panel bracing
~ air bag surcharging system in place
® 12 kPa surcharge applied
2.0

1.5

50 100 150 200 250

time from start of construction (hrs)

Figure 7 Propped Panel Wall Construction Sequence

Table 3

Surcharging Schedule for Propped Panel Wall

Event Elapsed Duration (hrs)

time (hrs)
wall construction
up to removal of bracing 0 217
air bag installation 217 25
12 kPa surcharge applied 242 430
30 kPa surcharge applied 672 100
40 kPa surcharge applied 772 110
50 kPa surcharge applied 882 1000
30 kPa surcharge applied 1882 16.5
12 kPa surcharge applied 1899 46
3.3 kPa surcharge applied 1945 43
excavation commences 1988 52
excavation complete 2240 -
82

III during construction

o at excavation
moisture
dry density ,Pd (kg/m 3 ) content, w (%)
3.5 -
.... ....t:
P'"
backfill 0 0 0 0 0 0
It) It) CD CD 0 0 0
layers t: ~ t: t: N It) .t-
o 0 0 I I I 0 I

3.0 A &
-
..
f-
4
"0 ~
,..-...
S
'--'
2.5 A 4
..0
cd
.-
rn
Q.)
rn 3
cd
..0
Q.)
2.0 l- • ..a ~ -
?-
0
..0
cd
-" ~

+J
r:::::
0
......
cd
1.5 ~
2
-
?-
Q.)
.-
A 0&
Q.)
.-
.- ....
~ 1.0

1
0.5 ~ 6 ~ -
-
A

o ~ ~ ~ I _I _I ..1- ~

a) Incremental Facing Wall Construction

Figure 8 Sand Fill Density and Moisture Content Profiles

 83

~ during construction

o at excavation
moisture
dry densitY,Pd(kg/m 3 ) content, w (%)
3.5 ,..
•t::
0 0 0 0 0 0
", CD f"- CD C! 0 0
....
It)
t: f"- f"- f"- N ~ ~
I I I I

3.0 !"

~---- --- ~ -C>--- - -04--- -

~

S 2.5
'--'
- • .t. ..
-
..D
cd
rJl

..
- -- --e----
Q)
rJl
cd 2.0 ~---- ~~---
..D

•
..
Q)
> .t.
0
..D
cd
~
0
1.5 -
---- - -
.~

--
+"
cd 1 - - - _ . ~-- -~---
>
Q)
Q)

te:
1.0 ~
• A

0.5 - - ---I --- A ~-- - --e- ---

A

•
o I I . • •

b) Propped Facing Wall Construction

Figure 8 (cont'd) Sand Fill Density and Moisture Content Profiles

84

5.0 Instrumentation

5.1 General
Each wall was instrumented to record the following:
a) horizontal movement of facing panels,
b) longitudinal displacements and strain in the geogrid reinforcement,
c) load at the panel/geogrid connections,
d) distribution of vertical earth pressure below the reinforced block of soil,
e) vertical settlements at the top of surcharge and,
f) temperature in fill.
The instrumentation used to record the data identified above was installed as
construction proceeded and was monitored for the duration of testing including wall
excavation.

5.2 Horizontal Movement of Facing Panels

In each test, horizontal movements of the central facing panels were monitored by
an array of electrical displacement measuring devices (potentiometer type).

5.3 Longitudinal Displacements and Strains in the Geogrid

5.3.1 General
The displacement of the reinforcement in the longitudinal direction of the model
walls was monitored by extensometers (i.e. by attaching tensioned steel wires to se-
lected locations on the mesh). Movement of the thin steel wires was recorded by
electronic displacement transducers mounted at the back of the test facility. The wires
were protected from the granular fill by passing them through stiff plastic tubes. In the
incremental wall construction, two displacement monitoring points per reinforcement
layer were used. In the propped wall test up to six displacement monitoring points
were used for each strip.
In addition, strain gauges were attached at selected mid-rib locations along the
length of each central reinforcement strip. After several years of experimentation with
different techniques of geogrid strain-gauging, the authors have found that high-strain,
foil-type gauges manufactured by Showa Measuring Instruments Co., Ltd. perform
satisfactorily. These strain gauges (together with proper rib surface preparation and
a two-part RTC epoxy adhesive) are effective at measuring small levels of strain in
the polymer reinforcement while the extensometers are effective at determining large
strains after the strain gauges have failed.
5.3.2 Strain Gauge Calibration
Experience with a variety of TENSAR geogrid reinforcement products has shown
that, in general, strain distribution is not uniform along any reinforcement rib. This
non-uniformity is a consequence of grid geometry and variable material moduli. In
order to make comparable strain measurements between strain gauge readings and
strains deduced from displacements recorded over several grid apertures it was neces-
sary to correlate strain gauge readings against gross average strains in the grid. In this
paper, strains measured over a gauge length consisting of one or more apertures are
called grid strains and strains measured with a gauge mounted at the mid-point on a
rib are called rib strains. It can also be appreciated that grid strain is a more meaning-
ful parameter since it facilitates comparison of mechanical response between different
sheet reinforcement types and is more easily implemented in analytical models.
 85

Recent investigations at RMC show that the relationship between grid and rib
strain can vary with geogrid type (e.g. Jarrett and Bathurst, 1987). In the current
investigation the correlation between strain gauge readings giving rib strain and grid
strain was determined from load-controlled in-isolation tests carried out on sections of
SR2 geogrid. A series of increasing static loads was applied rapidly to test specimens
and each load increment was sustained for 24 hrs while strain gauges and grid strains
were recorded electronically. The results of this test program are summarized on Figure
9. For SR2 material in the longitudinal direction, the data shows that grid strains and
rib strains are essentially equivalent up to about 1.8% grid strain. The strain gauges in
this test were arranged over two rows of ribs. Nevertheless, there was some deviation
about the average strain gauge reading at any given grid strain. This variability is
thought to be due to the sensitivity of gauge response to small differences in gauge
positioning, rib dimensions and non-uniform load distribution between ribs.

5.4 Load at Panel/Geogrid Connections

The facing panel/ geogrid connections for the incremental panel wall were con-
structed by clamping the geogrid between a horizontal steel-reinforced wood batten
and the back of the facing panels. Each batten was secured to the timber plywood
facing panels by four aluminium bolts passing through the panel.
For the propped wall construction the connection detail was modified so that the
geogrid reinforcement and panel bolts were at the same elevation. This configuration
proved to be a more secure method of attaching the geogrid to the facing panels.
Each of the four aluminium bolts in the panel/ geogrid connections was used as a
load cell by attaching four strain gauges to a reduced cross-section of the bolt. The
strain gauges were monitored to determine (horizontal) panel/reinforcement connection
loads at each reinforcement layer.

5.5 Distribution of Vertical Earth Pressures

In both tests a series of earth pressures cells was located at the bottom of the
test facility to record the magnitude and distribution of vertical overburden pressures
below the reinforced block of soil. The location of these devices is shown on Figures 3
and 4.
For the incremental wall construction, a series of 130 mm diameter aluminium
diaphragm-type pressure cells with an aspect ratio of 0.1 were manufactured in-house.
These cells proved to be unsatisfactory. For the propped wall, larger (230 mm diameter)
custom-built GEOKON earth pressure cells were used. These devices are 13 mm thick
(aspect ratio 0.06) and are of the cavity diaphragm-type with the cavity filled with an
incompressible liquid. Fluid pressures in equilibrium with vertical earth pressures are
measured by pressure transducers. These devices appeared to give a more accurate
measure of earth pressures.

5.6 Additional Instrumentation and Data Acquisition

Additional instrumentation included three displacement transducers which were
installed to measure vertical deformations at the top of the reinforced soil. Lastly,
temperature gauges were buried in the soil. This instrumentation was included to
ensure that measured ambient soil temperatures were close to 20°C corresponding to
the standard laboratory temperature for in-isolation tests carried out at the Netlon
laboratories and at RMC on SR2 reinforcement samples.
The instrumentation described in the previous sections was monitored using a
Hewlett Packard 3497A data acquisition system with an HP Vectra AT micro-computer
as the controller. The system allowed us to efficiently monitor up to 300 electrical
devices and provided fully-integrated data processing.
86

t In-Isolation Test Apparatus

clamp
SR2 TENSAR Geogrid sample
I I I I I (longitudinal direction)
15 ribs wide by 5 bars long
I
"
clamp
I I
10 strain gauges mounted at
mid-rib locations

2.0r--------------,
NOTES:
6 static load increments
load increment duration = 24 hrs
,--..
maximum load increment = 10 kN/m
~ temperature = 20 ± 2°C

general range

o 2.0
grid strain (%)

Figure 9 Calibration of Strain Gauges to Tensar SR2 Grid Strain

using Static Load Increment In-isolation Tensile Test
 87

6.0 Test Results

6.1 General
The following sections summarize the results of measurements taken over the
course of each test including wall construction, surcharging, and wall excavation.
6.2 Incremental Panel Facing Wall

6.2.1 Horizontal Panel Movements

Figure 10 shows panel movements recorded by the four displacement potentiome-
ters attached to the second panel of the incremental wall. The results of similar move-
ments recorded for all panels have been combined to produce Figure 11.
Figure 11 shows that the position of the top of the wall at the the end of the
surcharging program was about 40 mm from the initial position of the bottom panel.
However, about 14 mm of this movement was accumulated prior to placement of the
top panel. From Figure 10 it can be seen that the largest outward movements were
recorded as the surcharging loads were applied. However, at constant surcharge loads,
time-dependent outward panel movements were also apparent. For example, about 2
mm of outward movement in the top panel can be attributed to creep in the reinforced
wall model between 100 hrs and 500 hrs of the 50 kPa surcharging increment (Figure
11). Figure 11 also illustrates that progressive rotation of each facing unit about its
base occurred over the course of testing and that net rotation of the full wall height
about the toe of the structure was evident.

6.2.2 Geogrid Displacements

The incremental wall model was instrumented with two extensometers attached
to each reinforcement layer. The horizontal displacements recorded by these devices at
selected times during and after wall construction are given on Figure 12. Also shown
on the figure are outward panel movements recorded at the end of the 50 kPa sur-
charging increment. The figure shows that horizontal outward translation measured at
the geogrid 160 mm behind the panel facings was less than the magnitude of outward
panel movements recorded at the reinforcement elevations. However, the difference
can be attributed to strain in the grid, some play in the extensometers and compliance
in the panel/geogrid connections and the panels themselves. Nevertheless, qualitative
features of the geogrid movements measured immediately behind the panel units are
similar to those reported for the panels themselves on Figure 11. In particular, hori-
zontal deformations increase with height of reinforcement in the compdsite section and
with the magnitude of surcharge loading.
The data on Figure 12 indicates that there was significant horizontal translation
of the grid at distances up to a monitored distance of 1260 mm behind each panel.
Furthermore, the figure shows that, between 160 mm and 1260 mm behind each panel,
grid displacements were progressively attenuated indicating that tensioning of the grid
took place with increasing surcharge level.
6.2.3 Geogrid Strains
Selected strain gauge readings with time for the incremental panel wall test are
given on Figure 13. Some variability in strain gauge readings is observed for gauges
located at nominally equivalent distances behind the facing panels on the same rein-
forcement strip. This variability is likely due to non-uniform load distribution across
the reinforcement width, local variability in sand/geogrid interlock and small devia-
tions in details of strain gauge application. Nevertheless, all gauges on a given row
showed similar trends and grid strain distribution in the longitudinal direction of the
reinforcement is considered to be well-represented by the average strain gauge reading
at each row.
88

surcharge

TOP displacement
potentiometers (2)
3m
PANEL 2

BOTTOM displacement
potentiometers(2}

6m

14

S 12
5
10

6
~------
BOTTOM

~/--40
4 j.50kPa
kPa
2
r-12 kPa surcharge 30 kPa

4000
elapsed time (hrs)

Figure 10 Panel 2 Movement versus Elapsed Time

(Incremental Panel Wall Test)
 89

{i) Panel position just prior to bracing removal

@ Panel position after movement due to panel
o
#1 bracing removal

o@
Panel position after movement due to panel #2 bracing removal
Panel position after movement due to panel #3 bracing removal
Panel position 40 min after panel #4 bracing removal
@ Panel position after 12 kPa surcharge for 100 hrs
o Panel position after 12 kPa surcharge for 1000 hrs
@ Panel position after 30 kPa surcharge for 100 hrs
® Panel position after 40 kPa surcharge for 100 hrs
(@) Panel position after 50 kPa surcharge for 100 hrs
® Panel position after 50 kPa surcharge for 500 hrs
3.oor-----T5-----TIo---,,,15~_,~2°r_--172r5----~3TO~__~~~~4,0

geogrid LAYER 4
2.5

I 2.0

~ geogrid LAYER 3
EbO
- - - - -
"
'6
~
.D UI

"0>
.D

1:
'"
bO
'OJ geogrid LAYER 2
"" 1.0

geogrid LAYER 1

10 15 20 25 30 35 40

facing panel position (mm)

Figure 11 Summary of Panel Positions (Incremental Panel Wall Test)

90

CD 12 kPa for 100 hrs

® 12 kPa for 1000 hrs
12 ® 30 kPa for 100 hrs
G) 40 kPa for 100 hrs
@ 50 kPa for 100 hrs
® 50 kPa for 500 hrs
4 ~ 0 Panel movement after 50 kPa surcharge for 500 hrs

~ geogrid LAYER 4
°0~----------~1.~0-----------2~.~0----------~3.0

geogrid LAYER 3
1.0 2.0 3.0

~ ","grid LAYER'
°0~----------~1.~0-----------2~.0~--------~3.0

4 l !_ _ _-,.,_-o-.n. geogrid LAYER 1

,
00 1.0 2.0 3.0

distance from wall panel (m)

Figure 12 Summary of Geogrid Movements (Incremental Panel Wall Test)

 91

surcharge

LAYER 3, ROW 3
strain gauges

3m (3 gauges per row)

6m

1.0~------------------------------------------~

0.8

r
right

~ 0.6
'£...-=a 12 kPa ,.,w."g•
~
:9 0.4 \. 50 kPa
...
."..------ -------- J..-40 kPa

0.2 1---30 kPa

°0~~~----~IO~O~O~~~~2~OOO~~--L---=3~OO=O=---L---~4~000

elapsed time (hrs)

Figure 13 Typical Strain Gauge Response versus Elapsed Time

(Incremental Panel Wall Test)
92

Figure 13 also shows that sudden changes in strain gauge response match critical
events in wall construction and the application of increased surcharge loads. It is also
apparent from the figure that the polymer reinforcement continued to deform under
constant surcharge loading.
The average rib strains recorded along the length of each reinforcement layer are
summarized on Figure 14. The results of in-isolation tests described earlier indicate
that rib strains shown on the figures are essentially equivalent to grid strains. A number
of important observations can be made from these figures:

a) Significant strains were recorded in the reinforcement during construction of the

wall. The progressive development of strain in the reinforcement due to the in-
cremental wall construction can be noted from the data. The maximum strain
in the reinforcement layers at the end of construction decreases with increasing
reinforcement elevation. In layer 1, more than 50% of the total strain recorded at
the front of the reinforcement occurred during construction. '
b) Strain in the reinforcement is observed to increase with surcharge load level and
elapsed time under constant surcharge load.
c) In reinforcement layers 3 and 4 there is a marked peak in the longitudinal distri-
bution of strain at about 450 mm behind the facing panels. A similar trend could
not be established for layer 2 due to premature gauge failure.

Figure 15 shows a comparison of grid strains calculated from the extensometers

with grid strains inferred from rib-mounted strain gauges. The data shows that the
values from the displacement devices were never less than 70% of the average strain
gauge values taken over the same gauge length. The lower values can be attributed
to compliances within the extensometers. The relative values gave us confidence that
the reinforcement strains inferred from the strain gauges mounted directly on the
reinforcement were reasonable and that strains calculated from the extensometers were
lower-bound estimates of grid strains.

6.2.4 Load in Panel/Geogrid Connections

Strain gauges mounted on the bolts anchoring the paneljgeogrid connections at

the back of each facing panel were used to calculate horizontal load in these connections
at the end of the test (i.e. 50 kPa surcharge for 500 hrs). The results of these mea-
surements are plotted on Figure 16. Also shown on the figure are estimated connection
loads based on strains recorded in the first row of strain gauges. To calculate the loads
in the grid, isochronous load-strain-time data supplied for the actual reinforcement
material was used (see Appendix B). The method used to equate grid strain readings
to load has been reported by McGown et al. (1984) and Yeo (1985). Geogrid forces
calculated in this manner are considered to give upper-bound values on the connection
loads. Calculations using both methods to estimate connection loads give a range of
values between 0.5 and 4 kN jm at the end of the test with a 50 kPa surcharge. If the
maximum connection loads from the range of values shown on the figure are considered
then, the data indicates that the largest connection loads occur in the top and bottom
panels.
 93

geogrid LAYER 1
0.6

0.5
0 CD end of wall construction (bracing removed from LAYER 4)
® 12 kPa surcharge for 100 hrs
~0.4 ® o 50 kPa surcharge for 500 hrs (end of test)
~
'"
~
';;J 0.3
CD NOTE: strains are those recorded after end of construction
...
+'
(i.e. after panel bracing removed from LAYER 1)
"'
:g 0.2
0.25 m compacted backfill in LAYER 2

0.1

00 1.0 1.5 2.0 2.5 3.0

distance from panel wall (m)

NOTE: strains are those recorded after end of construction

(i.e. after panel bracing removed from LAYER 2)
geogrid LAYER 2
f7\ end of construction(bracing removed from
\.!..I LAYER 4)
® 12 kPa surcharge for 100 hrs
® 12 kPa surcharge for 1000 hrs
0.4 @ 30 kPa surcharge for 100 hrs
® @ 40 kPa surcharge for 100 hrs
ii
~0.3
~
o ®
(!)
50 kPa surcharge for 100 hrs
50 kPa surcharge for 500 hrs(end of test)
';;J
...
U; 0.2
.n
·c
0.1 ::::: ::::
::: :::: =
- == ===
~
~ II
°0~-------0~.~5--------~--------~------~~----~~~~~"~.0
1.0 1.5 2.0
distance from panel wall (m)

Figure 14 Summary of Strain Gauge Readings

(Incremental Panel Wall Test)
94

o
0.8 geogrid LAYER 3
end of wall construction(bracing removed from LAYER 4)
0.7 @ 12 kPa surcharge for 100 hrs
® 12 kPa surcharge for 1000 hrs
0.6 o 30 kPa surcharge for 100 hrs
@ 40 kPa surcharge for 100 hrs
_0.5
~
~
@
@ 50 kPa surcharge for 100 hrs
G) 50 kPa surcharge for 500 hrs( end of test)
,,- 0.4 @
.~
NOTE: strains are those recorded
t;
after end of construction
~ 0.3 ?8 (i.e. after panel bracing removed)

0.2
CD

0.1

00 0.5 1.0 1.5 2.0 2.5 .0

dist a nce from panel wall (m)

0.9

0.8

0.7 geogrid LAYER 4

- 0.6 @ 12 kPa surcharge for 100 hrs

~ ® 12 kPa surcharge for 1000 hrs
.~ 0;5 0
@
o 30 kPa surcharge for 100 hrs
@
~

'n 40 kPa surcharge for 100 hrs

@
~ o
.;:: 0.4 50 kPa surcharge for 100 hrs
50 kPa surcharge for 500 hrs( end of test)
0 .3
NOTE: strains are those recorded
after end of construction

ffi
(i.e. after panel bracing removed)
0.2

0.1 12 kPa surcharge applied

00 0.5 1.0 1.11 2.0 2 .5 3.0

distance from panel wall (m)

Figure 14 (Cont'd) Summary of Strain Gauge Readings

 95

geogrid LAYER 4
0.8 NOTE: all strains are those recorded at the end
of test after 500 hrs of 50 kPa surcharge

0.4

0.8

~
'"~
v
O~----~~----~~----~~--~~~--~~----~
0 0.5 1.0
- ........

be geogrid LAYER 2
.S
.:: 0.8
'Cd
'"'"
+"

0.4
o

~ strain from resistance strain gauges

0.8
0- -c average strain gauge strain
0-0 strain from wire connection/rear
potentiometer system

1.0 1.5 2.0

distance from panel wall (m)

Figure 15 Comparison of Grid Strains from Strain Gauge Readings

and Steel Wire Displacement Devices
(Incremental Panel Wall Test)
96

NOTE: loads all calculated from end of test data after 500 hrs
at 50 kPa surcharge pressure
1>--0 range of load from geogrid strain gauges 0.22 m from panel
cr--o range of load from connection bolt strain gauges
6 estimated load at panel from extrapolation of strain
3.0 distribution
....., ::J:I - - - geogrid LAYER 4
]: 2.5
.£:J
..::lrn
<I)
rn
2.0 - ~ - c.. - - - - - - geogrid LAYER 3
ro
.£:J
<I)
:>
1.5 estimated from strain gauges 0.66 m back from panel

=orE"- - - - - - -
0
~
>:::i
lY - geogrid LAYER 2
0
.~ 1.0
:>

-
<I)
Q)

ri3 0.5 ~ - 1:l:I_ - - _4- - geogrid LAYER 1

1.0 2.0 3.0 4.0 5.0 6.0

panel connection load (kN / m)

Figure 16 Geogrid/Panel Connection Loads from Incremental

Panel Wall Test

6.3 Propped Panel Facing Wall

6.3.1 Horizontal Panel Movements

Outward panel movements with time are plotted on Figure 17. These movements
are referenced to the end of construction corresponding to the panel geometry just
prior to bracing removal. Similar to the results of the incremental wall, the propped
panel wall showed large outward movements at the application of each surcharge load
increment and significant deformation under constant surcharge loading.
Figure 18 shows outward panel movements recorded at selected times during sur-
charging of the propped panel wall. The data shows a progressive rotation of the facing
unit about the toe with a total outward movement at the top of the wall of about 13
rom at the end of the test. As in the incremental panel wall test, the data shows
significant wall movements under constant surcharging load increments.
 97

displacement potentiometers
surcharge
!
TOP

3m MID-HEIGHT

BOTTOM

I.. ..I
6m

12

MID-HEIGHT

BOTTOM
2

°O~----------~5~O~O~----------710~O~O~----------~15~O~O--~

elapsed time since end of construction (hrs)

Figure 1'7 Panel Movement versus Elapsed Time (Propped Panel Wall Test)
98

NOTE: movements are those recorded after construction

(i.e. after the removal of panel bracing)

3 .0

CD after bracing removal

® 12 kPa for 1 hr
® 12 kPa for 100 hrs
o 12 kPa for 430 hrs
elevation ® 30 kPa for 100 hrs
g~grid LAYER 3
® 40 kPa for 100 hrs
o 50 kPa for 100 hrs
@ 50 kPa for 1000 hrs
elevation
geogrid LAYER 2

elevation
geogrid LAYER 1

4 6 8 10 12 14
panel movement (mm)

Figure 18 Summary of Panel Movements (Propped Panel Wall Test)

6.3.2 Geogrid Displacements

The propped panel wall was instrumented with up to six extensometers attached
to each reinforcement layer. The horizontal displacements recorded by these devices are
given on Figure 19. Also shown on the figure are outward panel movements recorded
at the end of the 50 kPa surcharging increment.
Figure 19 shows that there was essentially uniform displacement of the grid layers
after the initial 100 hrs of the 12 kPa surcharging increment. For example, in the top
layer this movement was about 2 mm towards the wall which is somewhat less than the
3.3 mm outward movement recorded at the panel facing opposite this reinforcement
layer (Figure 18). This observation suggests that there was some initial slip of the grid
as soil/grid interlock was established during the initial loading increment. At higher
surcharge loads the increased slope of the displacement profiles shows that tensile
straining in the reinforcement layers and load transfer between the grid and soil was
occurring.
 99

G) after bracing removal

® 12 kPa for 100 hrs
® 12 kPa for 430 hrs
@ 30 kPa for 100 hrs
8 ® 40 kPa for 100 hrs
® 50 kPa for 100 hrs
6 o 50 kPa for 1000 hrs
® panel movement after 50 kPa
surcharge for 1000 hrs

geogrid LAYER 4

0.5 1.0 1.5 2.0 2.5 3.0

NOTE:
movements are recorded after panel bracing removed

geogrid LAYER 3

0.5 1.0 1.5 2.0 2.5 3.0

2i~ geogrid LAYER 2

00 0.5 1.0 1.5 2.0 2.5 3.0

geogrid LAYER 1

1.0 1.5 2.0 2.5 3.0

distance from wall panel (m)

Figure 19 Summary of Geogrid Movements (Propped Panel Wall Test)

100

6.3.3 Geogrid Strains

Selected strain gauge readings with time for the propped wall test are given on
Figure 20. Similar to observations made for the incremental wall test, the strain gauges
showed increased tension in the grid as surcharge load increments were applied and
they also recorded time-dependent strain under constant surcharge loadings.
The distribution of rib strains at selected times in the propped wall test are plotted
on Figure 21. A number of important observations can be made from Figures 20 and
21:

a) The magnitude of strain recorded at a given distance behind the facing can be
seen to increase with time and surcharge load level.
b) The propagation of strain into the reinforcement can be seen to increase with time
and surcharge load level.
c) In geogrid layers 1, 2 and 3 there is a distinct change in strain gradient at about
0.45 to 0.65 m behind the wall.
The strain distributions for the propped wall show qualitative features in the im-
mediate vicinity of -the wall facing which are different from those recorded for the
incremental wall. In particular, layer 4 shows a progressive increase in geogrid strain
as the distance to the panel facing decreases. At the same location in the incremental
wall there were reduced strain values. Layers 2 and 3 show qualitative features that
fall between the strain profiles recorded for the incremental wall and layer 4 in the
propped wall. The difference in strain response is considered to be due to the rela-
tive vertical compllance in the two wall constructions. In the incremental wall test a
horizontal layer of compressible foam was placed between panel units. After fill com-
paction and surcharging, it was noted that the foam filler compressed leading to an
overall 8hortening of the full wall height by roughly 10 mm. This degree of freedom
was not available in the propped panel wall which was constructed as a single facing
unit. Consequently, the fill moved down with respect to the panel unit during fill
compaction and surcharging and in the process generated additional tensile loading
in the reinforcement close to the geogrid/panel connections. The relative soil/panel
movement can be expected to decrease with distance below the top of the wall and
this accounts for the observation that additional geogrid loading in the vicinity of the
panel connections appears to diminish with lower reinforcement elevation.
Grid strains were also calculated from the results of extenspmeters (i.e. from the
gradients plotted on Figure 19). The results ofthese calculations are given on Figure 22.
The figure shows that these calculated strains under-register the grid strains inferred
from the rib-mounted strain gauges . .Nevertheless, both sets of data show generally
consistent trends with the extensometers tending to give lower-bound estimates of grid
strain. Again, the discrepancy is thought to be due to mechanical compliances in the
steel wire systems.
6.3.4 Load in Panel/Geogrid Connections
A significant number of strain gauges attached to the panel connection bolts were
observed to have failed when the propped wall was excavated. The poor performance
of the strain gauged bolts did not allow panel/geogrid loads to be calculated with
any confidence using this method of instrumentation. Nevertheless, an upper-bound
estimate of panel loads can be determined from the strains recorded in the grid at
strain gauge locations closest to the facing unit. The results of this exercise are given on
Figure 23 and show that the load in the panel connections increases with reinforcement
elevation.
 101

surcharge

LAYER 3, ROW 1
strain gauges

3m (3 gauges per row)

6m

1.0r------------------------------------------------,
panel bracing removed
0.8
30 kFa
---.
~
40 kPa
'--' 0.6 right
~

...,...
.~

- - ---- ----
[/J

..0
.;:: 0.4 center

left
0.2

1500

elapsed time since end of construction (hrs)

Figure 20 Typical Strain Gauge Response versus Elapsed Time

(Propped Panel Wall Test)
102

geogrid LAYER 1 NOTE: strains are those recorded after end of construction
(i.e. after panel bracing removed)
0.5
(0 1 hr after bracing removal
0.4 ® 12 kPa surcharge for 100 hrs
o 12 kPa surcharge for 430 hrs
~
';;;' 0.3 o 30 kPa surcharge for 100 hrs
.;;;0:
....
® 40 kPa surcharge for 100 hrs
@
o
tl 0.2 50 kPa surcharge for 100 hrs
.D
.;::
50 kPa surcharge for 1000 hrs (end of test)
0.1

2.5 3.0

NOTE: strains are those recorded after end of construction

- - - (i.e. after panel bracing removed)
geogrid LAYER 2
0.5 (0 1 hr after bracing removal
® 12 kPa surcharge for 100 hrs
o
o
0.4 12 kPa surcharge for 430 hrs

~ 30 kPa surcharge for 100 hrs

.~ 0.3 ® 40 kPa surcharge for 100 hrs

....... .@ 50 kPa surcharge for 100 hrs

OJ

.;:: 0.2
.D o 50 kPa surcharge for 1000 hrs (end of test)

0.1
- --
--:::-=:
1.0 1.5 2.0
distance from panel wall (m)

Figure 21 Summary of Strain Gauge Readings (Propped Panel Wall Test)

 103

geogrid LAYER 3

0.7
NOTE:
strains are those recorded after end of construction
(i.e. after panel bracing removed)
0.6

1 hr after bracing removal

0.5 12 kPa surcharge for 100 hrs

~ CD 12 kPa surcharge for 430 hrs

IU 0.4 30 kPa surcharge for 100 hrs
"
.~
@ 40 kPa surcharge for 100 hrs
~
.0
0.3
@ 50 kPa surcharge for 100 hrs
.;::
50 kPa surcharge for 1000 hrs (end of test)
0.2 @

0.1
@
®
00 0.5 1.0 1.5 2.0 2.5 3.0
distance from panel wall (m)

0.9
®
0.8 NOTE: strains are those recorded after end of construction
---{i.e. after panel bracing removed)

0.7 (0 1 hr after bracing removal

® @ 12 kPa surcharge for 100 hrs

® 12 kPa surcharge for 430 hrs
~ 0.6

..-
o 30 kPa surcharge for 100 hrs
.~ 0.5 ® ® 40 kPa surcharge for 100 hrs
'0: @ 50 kPa surcharge for 100 hrs
®
.0
'C
0.4 @ 50 kPa surcharge for 1000 hrs (end of test)

0.3

0.2

0.1

0
--
--
0.5 1.0 1.:1 &.0 2.5 3.0
distance from panel wall (m)

Figure 21 (Cont'd) Summary of Strain Gauge Readings (Propped Panel Wall Test)
104

geogrid LAYER 4
0.8
NOTE: all strains are those recorded at the end
of test after 500 hrs of 50 kPa surcharge
0.4

I I I
- - --
I -'
0.5 1.0 1.5 2.0 2.5 3.0
geogrid LAYER 3
0.8

~0.4

00
geogrid LAYER 1
0.8 l:Jr---6 strain from resistance strain gauges
_ strain from wire connection/rear

~
potentiometer system

00 0.5
distance from panel wall (m)

Figure 22 Comparison of Grid Strains from Strain Gauge Readings

and Steel Wire Displacement Devices
(Propped Panel Wall Test)
 105

NOTE: loads all calculated from end of test data after

1000 hrs of 50 kPa surcharge pressure
o-a range of load from geogrid strain gauges 0.22 m from panel
o estimated load at panel from extrapolation of strain
distribution
3.0
- - - - - - - 0--0 - - - 0 - geogrid LAYER 4
]:2.5
.D
~
~ 2.0 - - -0---0- - - - - - - - geogrid LAYER 3
~
.D
Cl)

~ 1.5
.D
o:l
~
- - -0-00 - - - - - - - - - geogrid LAYER 2
o
.~ 1.0
:>
Cl)
Q)

~ 0.5 -00--0 - - - - - - - - - - geogrid LAYER 1

°o~--~~~~--~~--~--~~--~~--~
3.0 6.0 7.0
panel connection load (kN/m)
Figure 23 Geogrid/Panel Connection Loads from Propped Panel Wall Test

6.4 Additional Results from Both Tests

The results of earth pressure cell measurements taken at the bottom of the re-
inforced soil block in the incremental wall test were disappointing due to poor cell
sensitivity. The cell response problem was overcome in the subsequent propped wall
test by using different earth pressure cells with a larger diameter. The results of earth
pressure readings taken during the propped panel wall test at the end of construction
and at the end of surcharging are given on Figure 24. The data shown was arrived
at by calibrating each cell response insitu against the initial 2 m depth of fill placed
during wall construction.
In general, all pressure cells at the base of the test facility recorded vertical earth
pressures in the reinforced propped wall test that were below the value calculated from
soil self-weight and surcharge. Portions of this under-registration may be attributed
to pressure cell response, sidewall friction in the test facility and facing/backfill inter-
action for those cells c1oseo" t.o the front of the test facility. The influence of sidewall
friction was' discussed at lengih during the NATO Workshop. As a result of these
discussions, a large-scale unreirJorced propped wall test was undertaken to examine
the contribution of sidewall fr;~ 'ion and facing friction to stability of soil within the
RMC Test Facility. The results 01 this investigation are reported by Bathurst and Ben-
jamin in these proceedings and show that about 10 to 12% of vertical earth pressure
under-registration at the soil base may be due to sidewall friction. Futhermore, the
unreinforced test showed that the effective friction angle between the rigid facing panel
and backfill with the geogrid/panel connections in place is about 33 0 •
106

130

120

au = ,h + q
50 kPa surcharge for 1000 hrs

Q) ~ ......... ,
]
~
90 ( ' , ........
Q) I ~
I-<
::s
rn
I '\,
"~
rn
Q)
I-< I 50 kPa surcharge first applied}.
0..
t
....""§ 70
~
Q)
>
q. \ \ end of wall construction

-~-~~------
'cI' ........................
50 "'C>- _
------0

distance from panel facing (m)

Figure 24 Vertical Pressure Distributions from Propped Panel Wall Test

The results of temperature measurements taken within the reinforced soil mass
showed that the fill temperature never varied from 21 ± 2°C. As a result, the mechani-
cal properties of the SR2 geogrid material established from laboratory in-isolation tests
at 20°C are considered representative of the insitu condition.
Finally, it should be noted that an attempt was made to monitor vertical settle-
ment of the reinforced sand backfill using displacement devices mounted at the top
of the test facility. However, the installation of these devices proved difficult and no
useful data was recovered.
 107

7.0 Discussion of Test Results

7.1 General
The purpose of this paper is to report the results of two large-scale reinforced soil
walls constructed at RMC. The trial walls provided a starting point for discussions re-
lated to the general theme of the Workshop (Le. application of polymeric reinforcement
in retaining structures). This section discusses the RMC test facility and qualitative
features of the model tests which can assist the reader to interpret the test results.

7.2 Wall Stability

Both test walls proved to be stable during construction and during sustained
surcharging of up to 50 kPa magnitude. While time-dependent panel movements were
still underway at the completion of each test, the rate of deformation was diminishing
with time. The grid strains inferred from strain gauges showed that the SR2 was
well within the 10% strain performance limit recommended for this material (McGown
et al., 1984). Even if the maximum recorded logarithmic rate of strain at the end
of each test was to persist over an additional three logarithmic time cycles (i.e. 120
years) the reinforcement layers would still have adequate tensile capacity under the 50
kPa surcharge. It is clear from the comments made above that the laboratory walls
were subject to working load conditions rather than a condition close to the ultimate
surcharge capacity of the composite systems.
Wall stability calculations based on limiting equilibrium methods may not be ap-
propriate to determine working tensile loads in the reinforcement layers. Having said
this it is of interest to note that the locus of maximum geogrid strains in layers 3 and
4 from the incremental panel wall test suggests that (if surcharge loads could be in-
creased) a near-vertical failure surface through the top half of the composite soil mass
might be generated about 0.45 m back from the wall. This feature is qualitatively
similar to critical failure surfaces assumed in coherent gravity methods of wall design
(e.g. Tensar Corp., 1986) and by Juran and Schlosser (1978) for reinforced soil walls.
A similar locus of maximum geogrid strains is masked in the propped panel wall test
by the generation of additional geogrid strains close to the panel/ geogrid connections
resulting from the relative vertical movement between the sand and the panels during
fill placement and surcharging. While some features of the coherent gravity method
of design are observed in the data from the current study, both walls showed outward
rotation of the walls about the toe at working loads rather than rotation about the
top which is often assumed in limiting equilibrium-based coherent g:t:avity methods of
design. The results of these tests suggest that the condition of bottom restraint in
these walls is critical to the performance of these structures. In the current study, the
facing unit/levelling pad interface was sufficiently rough to offer an essentially fully
restrained horizontal and vertical support. A different deformation response would be
anticipated for walls constructed with a less restrained horizontal degree of freedom at
this boundary.
The RMC retaining wall test facility was constructed to act as a (near) plane-strain
apparatus. However, despite friction-reducing sidewall construction and separation of
the reinforced soil blocks into three sections, the sidewalls will contribute to the overall
stability of the model walls. The magnitude of this contribution is, however, difficult
to quantify based on measured results from the trial walls. In the companion paper by
Bathurst and Benjamin, the results of an unreinforad test using the same sand backfill
and propped wall construction are reported. The purpose of this test was to determine
the contribution of sidewall friction and facing/sand interaction to the stability of the
soil within the retaining wall test facility. The results of analyses using measured
boundary forces on the unreinforced wall and results of direct shear box tests on the
sidewall/sand interface suggests that for the propped reinforced wall approximately
14% of the resistance to active earth force may be due to sidewall friction (assuming
that limiting equilibrium conditions are applicable to the reinforced test).
108

The sand backfill employed in these large-scale tests has a high shear strength.
The results of direct shear tests show that the material has a peak friction angle
between 42° and 53° and possibly as high as 55 or 56° if plane-strain conditions are
considered (Jewell, 1987b). In addition, it was observed during excavation of earlier
wrap-around constructions that the residual moisture content in the sand gave the
material an apparent cohesion great enough to allow a cut face one meter high to
stand unsupported. The inherently high shear strength of the sand backfill is thought
to explain in part why relatively low tensile loads were generated in the reinforcing
elements.

7.3 Summary of Measured Results

As part of the Workshop, participants were asked to make Class A predictions of
the behaviour of the incremental and propped panel walls reported in this paper. A
summary of measured results is appended to this paper (Appendix C).

8.0 Conclusions
Two(2) large-scale model reinforced soil walls were constructed within the soil
retaining wall test facility at RMC. The walls comprised an incremental panel wall
construction and a single propped panel wall system. Both walls were constructed
with a Tensar SR2 geogrid as the sand backfill reinforcement. Each wall was subjected
to sustained surcharge loading up to 50 kPa following construction.
The following points summarize some of the important observations made during
the construction and surcharging of these structures.

1) Both large-scale model walls proved to be stable during construction and surcharg-
ing. The low levels of strain developed in the reinforcement rayers and decreasing
rates of panel movement suggest that both walls would have remained stable for
many years under sustained 50 kPa surcharging.
2) The relative vertical compressibility of the facing units in the two tests influenced
the generation of geogrid strains in the vicinity of the wall panels. In the incremen-
tal wall construction, the panel units were better able to settle with the retained
sand backfill owing to a layer of compressible foam placed at each horizontal joint
between panels. In the propped panel wall this degree of freedom was not present
and additional tensile geogrid straining was recorded in the vicinity of the ge-
ogrid/panel connections as the backfill settled under the action of fill self-weight
and surcharging. These observations clearly have important implications to the
design of these structures. Specifically, details of wall type and construction can
lead to additional geogrid loads which are not accounted for directly in current
design methodologies.
3) The history of panel deformation and strain development in the reinforcing layers
was sensitive to the construction technique employed. In the incremental wall,
significant tensile strains were developed during construction. In addition, larger
outward panel movements were recorded for the incremental panel wall than for
the propped wall test.
4) Estimated geogrid/panel connection loads were very low in these large-scale mod-
els. In the incremental wall construction, the trend in geogrid strains close to the
panels suggests that earth pressures on the facing units are very small due to the
reinforcing effect of the geogrid layers.
5) The distribution of vertical earth pressures below the reinforced soil block in the
propped wall test may be markedly reduced close to the wall facing indicating that
significant backfill/facing interaction occurs. A portion of the under-registration
recorded by pressure cells further from the rigid wall facing is considered to be
due to sidewall friction.
 109

Acknowledgements

The authors would like to acknowledge the collaboration of Drs. A. McGown and
K. Andrawes of the University of Strathclyde, Scotland in the conceptual development
and pursuit of this testing program. Appreciation is also extended to D. Varney and
D. Swan, graduate students at the University of Strathclyde for assisting in the con-
struction of the walls. Funding to support this liaison between RMC and Strathclyde
was provided through a NATO Collaborative Research Grant. The authors would also
like to thank Netlon U.K. for carrying out the in-isolation tests used to determine the
load-strain-time data for the SR2 reinforcement and for the provision of the material.
Additional support was provided by members of the Civil Engineering Dept. at RMC
including Messrs. J. Bell, D. Wawrychuk, S. Prunster and J. DiPietrantonio. The in-
isolation tests for strain gauge calibrations were performed by Capt. G. Richardson,
graduate student at RMC. Financial support for construction of the test facility and
reinforced soil wall models was provided through the ARP program and the Chief of
Construction and Properties, DND (Canada). Finally the authors would like to thank
Dr. G.W.E. Milligan for his many fruitful discussions while on a sabbatical visit from
Oxford University.
110

References

1. BATHURST, R.J. and JARRETT, P.M. (1986)

Class A Prediction Exercise for Reinforced Earth Walls,
Bulletin No.1 for NATO Advanced Research Workshop, Application
of Polymeric Reinforcement in Soil Retaining Structures
Departments of Civil Engineering RMC and University of Strathclyde
2. BATHURST, R.J. (1987)
Large-Scale Reinforced-Earth Wall Tests at RMC
Report to Chief of Construction and Properties
Dept. National Defence (Canada)
Department of Civil Engineering RMC
3. BATHURST, R.J. and BENJAMIN, D.J. (1987)
Preliminary Assessment of Sidewall Friction on Large-Scale
Wall Models in the RMC Test Facility
Application of Polymeric Reinforcement in Soil Retaining
Structures, Nato Advanced Research Workshop,
Royal Military College of Canada, June 1987
4. BATHURST, R.J. and KOERNER, R.M. (1987)
Results of Class A Predictions for the RMC Large-Scale
Reinforced Soil Wall Trials
Application of Polymeric Reinforcement in Soil Retaining
Structures, Nato Advanced Research Workshop,
Royal Military College of Canada, June 1987
5. DEPARTMENT OF TRANSPORT, U.K. (1978)
Reinforced Earth Retaining Walls and Bridge Abutments for
Embankments, Technical Memorandum BE3/78
6. JARRETT, P.M. and BATHURST, R.J. (1987)
Strain Development in Anchorage Zones
Proc. of Geosynthetics '87 Conference, New Orleans
7. JEWELL, R.A. (1987b)
Analysis and Predicted Behaviour for the RMC Trial Wall
Application of Polymeric Reinforcement in Soil Retaining
Structures, Nato Advanced Research Workshop,
Royal Military College of Canada, June 1987
8. JURAN, I. and SCHLOSSER, F. (1978)
Theoretical Analysis of Failure in Reinforced Earth Structures
Proc. Symp. on Earth Reinforcement, ASCE, Pittsburgh
9. LESCOUTRE, S.R. (1986)
The Development of a Large-Scale Test Facility for Reinforced
Soil Retaining Walls
M.Eng. thesis, Royal Military College of Canada, Kingston
10. McGOWN, A., ANDRAWES, K., YEO, K. and DUBOIS, D. (1984)
The Load-Strain-Time Behaviour of Tensar Geogrids
Symp. on Polymer Grid Reinforcement in Civil Engineering
Paper No. 1.2, London
11. TENSAR CORP. (1986)
Guidelines for the Design of Tensar Geogrid Reinforced Soil
Retaining Walls
 111

12. WAWRYCHUK, W.F. (1987)

Two Geogrid Reinforced Soil Retaining Walls
M.Eng. thesis, Royal Military College of Canada, Kingston
13. YEO, K.C. (1985)
The Behaviour of Polymeric Grids used for Soil Reinforcement
Ph.D. thesis, University of Strathclyde, Scotland
112

Appendix A

DIRECT SHEAR TESTS

1l1li
DENSITY. 1.69 Mg/m 3
II

II

70
""
c..
~
0;,= 67.7 kPa
81
en
en
lJJ
e: 51 54·2
I-
en 40.6
e: «I
-<
lJJ
:J: JI
en
--cr-_ _-----.....--~---...---..----27.'
2Q

__---------_13.5
II

o .5 1 1.5 2 2.5
HORIZONTAL DISPLACEMENT (mm)

DIRECT SHEAR TESTS

DENSITY - 1.69 Mg/m 3

~
I-
:z
lJJ
ffi
<.J
-<
.-J
c..
en
~
Cl
.-J
-<
<.J
~
l-
e:
lJJ
>

..II5-+-~-..,....----r----Y-----.---....,....---I
1 1.5 2
HORIZONTAL DISPLACEMENT (mm)

Figure A.la Load-Deformation of Sand Backfill Material

from Direct Shear Box Tests (Pd = 1.69Mg/m3 )
 113

DIRECT SHEAR TESTS

1111
DENSITY. 1.78 Mg/m J
!II

II

ro 70

=
Cl..

en
en
II

tJ.J
a: !III
~
en
a: c
<
tJ.J
:c !II ~_--".--------~---_ 27.1
en

211
__ ----------------_1~
10

0 .5 1 1.5 2 2.5
HORIZONTAL DISPLACEMENT (mm)

DIRECT SHEAR TESTS

. DENSITY. 1.78 Mg/m J

~
:z
UJ
:::L
UJ
u
<
...J
Cl..
en
......
CJ .1
...J
<
u
...... .1
~
a:
UJ .4
>
.2

o .5 1 1.5 2 2.5
HORIZONTAL DISPLACEMENT (mm)

Figure A.lb Load-Deformation of Sand Backfill Material

from Direct Shear Box Tests (Pd = 1.78Mgjm3 )
 114

DIRECT SHEAR TESTS

DENSITY ~ 1.86 Mg/m 3
II

;;
C.
e CY,,: 67.7 kPa
en
en
UJ
a::
I-
en 54.2
a::
-<
UJ
:x:
en
2D 27.1

10 13.5

0 2 3 4 6
HORIZONTAL DISPLACEMENT (mm)

DIRECT SHEAR TESTS

. DENSITY ~ 1.86 Mg/m 3
u

e 3.2
..s
I- 2.
Z
UJ
::&:
UJ 2.4
U
-<
....J

....fk
Cl
1.1
....J
-<
....
U
I-
a::
UJ
>
.4

3 4 6 8 9
HORIZONTAL DISPLACEMENT (mm)

Figure A.1c Load-Deformation of Sand Backfill Material

from Direct Shear Box Tests (Pd = 1.86Mg/m3 )
 115

80

10

Shear Box size = 36 cm 2

60

50
.~
0...
C-
oo
40
....,...
00
Q)

...
00

~ '"
Q)

30
'"

20 1.69

0---0 1.78

10 0----0 1.86

1.86 40° (const vol)

eo
normal stress (kPa)

Figure A.2 Mohr-Coulomb Failure Envelope for Sand Backfill Material

from Direct Shear Box Tests
116

260

240

220

200

...... 180
0
a.
.>I!

160
W
u
z
W
a:: 140
W
u.
!:
0 120
(J)
(J)
W
a:: 100
I-
(J)

------=-
(7"3=30 kPo
..J
80
~
u 0"3= 30 kPo
~
a:: 60
a.

40

20

0
2 4 6 8 10 12 14 16 18 20
-I
~~-2
et:~

I- -3
Wz
~-
:::::><1: AXIAL STRAIN (%l
..Jet:
01-
>(J)

Figure A.3 Stress-Strain Behaviour of Sand Backfill Material

from Standard Consolidated-Drained Triaxial Compression Tests
(Pd = 1.68Mg/m3)
 117

180

- 160
~
.x
- 140
....
.; 120
VI
(7"3 =60kPa
ILl
~ 100
VI

0:: 80
<t
ILl
~ 60

40

20

40 80 120 160 200 240 280 320

NORMAL STRESS. (7" (kPa)

Figure A.4 Mohr-Coulomb Failure Envelope for Sand Backfill Material

from Standard Consolidated-Drained Triaxial Compression Tests
(Pd = 1.68Mgjm3 )
118

Appendix B

80r-----1I----~------~----_r----~~-- ___

70

50

40

30 v PEAI< LOAD

20

10

o
o 5 10 15 20 25 30
STRAIN (%)

Figure B.l

Constant Rate of Strain Test, Load-Extension for TENSAR SR2 at 20°C

 119

ClEEP ST•• I (I)

5 II 15 21
I+-------~-----+------~----~

-1 +-----~1_--~---~------_+------~

-2+---~--~--~~rr------~------~

.
I~ ; -]
I z:
~
+----t--+-t----J~_4k_------_+------~

39·6

..• -
~ g§
CIt ...
! ~ -1 +-----+-+---+----'\+-------_+---------1
~

-5 .~--~~---r~~-------T-------~
13·2 ~ 28·7

-,+-------4--------r------~-----

-l~------~------~------~----~
LCIfDt ECPiEMlI
I" 1IIt/44 r It.
"IDIN
TEST TEMPERATlH: 20 t1°C 0

Figure B.2 Sherby-Dorn Plot for TENSAR SR2 at 20°C

120

45~--~----.-----r---~----.-----r---~----~

40 100h

35

30

-E
~ 25
o
~
oJ
20

15

10

2 6 8 10 11 14 16
STRAIN (0/0)

Figure B.3 Isochronous Load-Strain Curves for TENSAR SR2 at 20°C

 121

5
z
«
f!
VI
4
0~
"-
e
"-
z
-3
,x

VI
10% strain

VI
~
u..
u..
......
t-
VI 2
VI
::>
0
z
0
cr;
~

-
u 1
~
a..
l.U
l.U
cr;
U
0
100 10' 10a 10) 10" 105 10·
TIME (hours) [Log Scale]

Figure B.4 Creep Isochronous Stiffness versus Log Time Relationship

for TENSAR SR2 at 20°C
122

50

e 4{)
.
10 ( , " ..
,.
~
."

'-
z .-"

--~
~
-"
}) /
Cl ./
/
-I
/
20 /
/
I..
'10

00 2 4 6 8 10 12 14 16 18
STRAIN (%)

1'4

1'2
0:::
...
1,0
0:::
0
.....
u
Lf 0·8
z
....
0
..... 0-6
u
UJ
0:::
0:::
0
u
0"4
W'
.....
:I:
..... 0·2

0 1)1
10° '()' 1)3 1)" 10- 10·
TIME (hours) [Log Scal~]

Figure B.5 100 Hour Isochronous Curve and Time Correction Factor
versus Log Time Relationship for TENSAR SR2
 123

APPENDIX C

TABLE C.I

Horizontal Panel Positions at Reinforcement Elevations

<Incremental Panel Wall Test)*

surcharge= OkPa 12kPa 12kPa 12kPa 30kPa 40kPa 50kPa 50kPa

duration= ** 100hrs 1000hrs 1600hrs 100hrs 100hrs 100hrs 500hrs
<nun) <nun) <nun) <nun) <nun) <nun) <nun) <nun) .

layer 1 0 4 4 4 4 5 5 5

layer 2 6 9 10 11 13 14 14 15

layer 3 11 13 13 16 18 19 20 20

layer 4 14 18 18 21 24 25 26 27

* refer to Figure 3 for wall geometry

** panel positions at time of installation, post-construction move-
ments can be calculated by subtracting first column of numbers
from subsequent measurements

TABLE C.2

Horizontal Panel Movement at Reinforcement Elevations

<Propped Panel Wall Test)*

surcharge= 12kPa 12kPa 30kPa 40kPa 50kPa 50kPa

duration= 100hrs 430hrs 100hrs 100hrs 100hrs 100hrs
<nun) <nun) <nun) <nun) <nun) <nun)

layer 1 1 1 1 1 1 2

layer 2 1 1 2 2 3 4

layer 3 2 2 3 4 5 7

layer 4 3 4 6 7 8 10

* refer to Figure 4 for wall geometry

124

TABLE C.3

Strain in Reinforcement (Incremental Panel Wall Test)*

surcharge
duration
strain (%)

layer location** 12kPa 12kPa 12kPa 30kPa 40kPa 50kPa 50kPa

(mrn) 100hrs 1000hrs 1600hrs 100hrs 100hrs 100hrs 500hrs

1 220 0.41 0.42 0.42 0.47 0.50 0.52 0.52

445 0.07 0.04 0.04 0.06 0.07 0.08 0.07
675 0.03 0 0 0.01 0.01 0.01 0.01

2 220 0.38 0.39 - - - - -

445 0.22 0.25 0.26 - - - -
675 0.16 0.15 0.15 0.25 0.29 0.34 -
900 0.10 0.09 0.09 0.15 0.17 0.20 0.23
1350 0.06 0.04 0.04 0.07 0.09 0.10 0.11

3 220 0.31 0.34 0.33 0.41 0.45 0.49 0.49

445 0.37 0.39 0.39 0.57 0.66 0.75 0.79
675 0.29 0.31 0.31 0.46 0.56 0.65 0.70
900 0.22 0.24 0.24 0.34 0.39 0.47 0.50
1350 0.17 0.17 0.17 0.24 0.27 0.29 0.31
1900 0.25 0.25 0.25 0.33 0.35 0.36 0.37

4 220 0.18 0.18 . 0.17 0.37 0.41 0.49 0.50

445 0.22 0.25 0.25 0.53 0.64 0.76 0.89
675 0.18 0.20 0.20 0.38 0.46 0.55 0.58
900 0.19 0.20 0.21 0.28 0.37 0.46 0.49
1350 0.18 0.17 0.17 0.17 0.20 0.23 0.25
1900 0.17 0.14 0.13 0.11 0.14 0.18 0.20

* strains recorded after each panel support released

** refer to Figure 3 for wall geometry and strain gauge locations
denotes gauge failure
 125

TABLE C.4

Strain in Reinforcement (Propped Panel Wall Test)

surcharge
duration
strain (%)

layer location* 12kPa 12kPa 30kPa 40kPa 50kPa 50kPa

(mm) 100hrs 430hrs 100hrs 100hrs 100hrs 1000hrs

1 220 0.04 0.06 0.11 0.15 0.18 0.25

445 0.04 0.12 0.15 0.18 0.20 0.36
675 0.01 0.03 0.03 0.04 0.04 0.06

2 220 0.07 0.08 0.17 0.23 0.31 0.37

445 0.06 0.08 0.15 0.21 0.28 0.35
675 0.04 0.09 0.15 0.20 0.26 0.36
900 0.03 0.05 0.09 0.12 0.15 0.21
1350 0.01 0.04 0.06 0.07 0.09 0.11

3 220 0.06 0.12 0.21 0.28 0.36 0.44

445 0.05 0.08 0.15 0.22 0.29 0.35
675 0.07 0.16 0.28 0.37 0.47 0.67
900 0.06 0.13 0.22 0.29 0.37 0.57
1350 0.02 0.07 0.11 0.14 0.18 0.22
1900 0.01 0.05 0.07 0.10 0.12 0.15

4 220 0.10 0.22 0.39 0.52 0.66 0.88

445 0.05 0.13 0.26 0.35 0.46 0.55
675 0.05 0.08 0.18 0.25 0.33 0.40
900 0.04 0.07 0.15 0.21 0 . 28 0.36
1350 0.04 0.08 0.11 0.14 0.18 0.25
1900 0.03 0.06 0.11 0.15 0.19 0.23

* refer to Figure 4 for wall geometry and strain gauge

locations
RESULTS OF CLASS A PREDICTIONS FOR THE RMC REINFORCED SOIL WALL TRIALS

RICHARD J. BATHURST, Royal Military College of Canada

ROBERT M. KOERNER, Drexel University, U.S.A.

1. INTRODUCTION

The NATO Advanced Workshop on Application of Polymeric Reinforce-

ment in Soil Retaining Structures was undertaken to critically assess the
state-of-the-art in polymeric reinforced soil retaining structures. In
order to establish common ground for discussions, a number of participants
were asked to make detailed Class A predictions of the performance of two
large-scale model polymeric reinforced soil walls constructed within the
RMC Retaining Wall Test Facility.

A description of the two tests and the test results are presented
in the companion paper in this volume by Bathurst, Wawrychuk and Jarrett.

Participants were issued with details of the proposed walls "in

advance" of construction (Bathurst and Jarrett, 1986). The data included:

a) description of the RMC Retaining Wall Test Facility;

b) proposed construction sequence for each wall;

c) construction drawings for proposed walls;

d) material properties for the retained soil;

e) load-strain-time properties of the polymeric rein~orcement; and

f) surcharge loading schedule.

Participants were asked to predict the performance of the two walls

at various elapsed times after initial construction and after surcharging
loads had been applied to the top of the retained soil. Predictions were
to include the deformations and strains incurred as a result of the two
different construction techniques employed and time-dependent behaviour of
the polymeric reinforcement.

The following predictions were requested at the times shown in

Table 1:

a) horizontal panel movements at the reinforcement elevations;

b) strain in the reinforcement;

c) total load in the panel/reinforcement connections; and

127

P. M. Jarrett andA. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 127-171.
© 1988 by Kluwer Academic Publishers.
128

d) vertical pressure along the base of the reinforced soil mass.

Surcharge Load Duration

(kPa} (hrs}
12 100
12 1000
30 100
40 100
50 100
50 1000

Table 1
Schedule of Prediction Times
for Proposed RMC Trial Walls

In addition, the predictors were asked to estimate the magnitude of

the uniformly distributed surcharge load required to initiate wall fail-
ure.

Along with each set of predictions the predictors were asked to

submit a brief description of the method of analysis that they adopted to
arrive at their estimates. These submissions form a large part of the
current paper.

In keeping with the informal spirit of the workshop and to encour-

age as many predictions as possible it was decided not to attribute
predictions to their authors in the published proceedings.

2. RMC TEST WALLS

The large-scale reinforced soil test walls comprised:

a) an incremental panel construction; and

b) a propped (or tilt-up) facing construction.

The general arrangement for the two test wall configurations is

given on Figures 1 and 2. Both walls used a medium to coarse grained sand
as the soil backf ill and an SR2 Tensar "Geogrid" as the re inforcement.
The facing panels were constructed as timber bulkheads.

3. METHODS OF ANALYSIS

A total of 10 predictors submitted estimates of wall performance.

In some cases, only one of the walls was attempted and in others only a
portion of the prediction exercise was attempted. In several cases the
submissions were made by a team of workers. It should be noted that the
predictors have a wide variety of experience. Some are researchers at
universities and government laboratories, others are practicing geotech-
nical engineers and many are both. Each predictor or team adopted a
different approach to arrive at their best estimate of wall performance.
In light of these comments it is fair to say that the predictions repre-
sent a wide range of current strategies that can be employed to estimate
the performance of polymeric reinforced soil walls.
 129

SAND/AIR PRESSURE BAG SURCHARZE 12 . 4m~

UNREINFORCED, 7 SR2 GEOGRIOe:n,:-7 SURCHARGE
SAND BACKFIL~L_ REINFORCED ~ PANELS ~
----1--------(1--- e-
Ii \
- -S~AY;; \: :
T•
SR2 LAYER 3

SR2 LAYER 2

~
~
SR2 LAYER I

~l~D~N!.. ~y~ ~

J
• STRAIN GAUGE
L COOC" ETE LEVELLING PAD ~
_ PRESSURE CELL

CROSS - SECTION VIEW FRONT VIEW

Figure 1 General Arrangement for Incremental Wall Test

_ : AND/AIR PRESSURE BAG SURCHARlGE r- 2.4m--j

UNREINFQRCED
SAND BACKFILL /
7 SR2 GEOGRID ~_-::7
REINFORCED ~
SURCHARGE
PANELS A
•....

e- IJ \ r

SR2 LAYER 3 -...: '"w

~

z
1f:
SR2 LAYER 2
•"
o z
,.; 0
i:!

~
L J
• STRAIN GAUGE
ET • LEVELLING PAO ..:,.

_ PRESSURE CELL

CROSS -SECTION VIEW FRONT VIEW

Figure 2 General Arrangement for Propped (Tilt Up) Wall Test

4. PREDICTION METHODOLOGIES

4.1 General

The following descriptions are design method summaries submitted by

the predictors prior to the Workshop in June 1987. These overviews can be
used to compare essential features of the models employed by the
participants.
130

4.2 Methodology Summaries

PREDICTOR A

Al . PHILOSOPHY OF METHOD

The predictions are based upon a mathematical simulation of the

construction and loading procedures used to build the two walls. The
mathematical soil model used is based upon the hyperbolic equation pro-
posed by Konder (1963) and described by Duncan and Chang (1970) in which:

(0, - 0.) = _e:__ (AI)

a + be:

where 0, and 0. major and minor principal stresses.

and e: axial strain

a and b constants

By expressing the, parameters a and b in terms of the initial tangent

modulus value, the compressive strength equation (AI) may be expressed as:

e:
(0, - 0.) = (A2)
e:R f
L +
E. (0, - o')f
~

where (0, - 0.) f compressive strength or stress difference

at failure

initial tangent modulus

failure ratio defined by:

and (0, - o')ult the asymptotic value of stress difference

A2. MODELLING OF SYSTEM

Simulations of the construction and loading for the two walls are
shown on Figures Al and A2.

A3. PARAMETERS USED IN ANALYSIS

A3.1 , Fill

The fill was assumed to be a non-linear elastic material. No

unloading-reloading was assumed, and Poisson's ratio was taken as constant
(v = 0.35). Stress-strain behaviour of the sand backf ill was taken from
standard Consolidated-Drained Triaxial Compression Tests provided.
 131

, ,,
t
i J ! I /4 STEP 1. Elastic Base + weight of Step L

STEP 2. Plus weight of Step 3

STEP 3. Plus weight of Step 4

STEP 4. Plus weight of Step 5

STEP 5. Plus weight of Surcharge

STEP 6. Increase surcharge 12 - 30 kN/m 2

STEP 7. Increase surcharge 30 - 40 kN/m 2
STEP 8. Increase surcharge 40 - 50 kN/m 2

Figure Al Simulation of Construction and Loading of Incremental Wall

132

r I I
• r• •
Z
..
r I
.,
,
I
t
't*

STEP 2 . Plus weight of Step 3

STEP 1. Elastic Base + weight of Step 2

STEP 3. Plus weight of Step 4

STEP 4. Plus weight of Step 5

STEP 5. Plus weight of surcharge

2
STEP 6. Increase surcharge 12 - 30 kN/m
STEP 7. Increase surcharge 30 - 40 kN/m 2
STEP 8. Increase surcharge 40 - 50 kN/m 2

Figure A2 Simulation of Construction and Loading of Propped Wall

 133

A3.2 Compressive Stress

A compaction stress of 3.4 kN/m 2 was assumed for the reinforced

soil fill. This was determined by the use of the empirical formula:

a I =
h

where P = effective weight of the vibrating tamper.

A3.3 Reinforcement

An isochronous stiffness of 3.6 kN/m was assumed for the initial

100 hour analysis. A reduction of 10% in stiffness was assumed for the
1000 hour condition.

A3.4 Facing

The facing was assumed to have the same stiffness as the concrete
footing and to be linear elastic.

0.24 X 10· kN/m 2

Econcrete

Vconcrete = 0.17

A3.5 Fill Immediately Behind Facing

The layer of fill immediately behind the facing was assumed to be

elastic and to have a low modulus but high Poisson's ratio.

E 0.1 X 10 3 kN/m 2

v 0.49

A3.6 Fill Beneath Reinforced Soil Block

The layer of fill beneath the reinforced soil block was assumed to
be elastic.

E 0.35 X 10' kN/m 2

v = 0.35

A3.7 Overburden/Surcharge

The overburden was assumed-to be elastic.

E 0.1 X 10' kN/m 2

v = 0.49
134

A4. COMMENTS ON PREDICTION

A4.l General

Negative movements resulting from numerical simulations represent

movement away from the fill. Positive values represent movement into the
fill. The Horizontal Panel movements are very susceptible to the quality
of the facing/reinforcement connection. This prediction assumes complete
fixity. In reality a large degree of slackness can be expected, in which
case the movement of the facing will be away from the fill (Le. this
prediction is expected to fail at this point).

A4.2 Propped Panel Wall

In this trial wall, the connection of the reinforcement, to the

facing makes no allowance for vertical movements. Hence, this structure
does not follow one of the recognized methods of construction (concertina,
telescope or sliding). The case can be classified as a 'special' propped
wall. Some level of increased stress at the reinforcement/facing
connections can be. expected.

REFERENCES

1. Duncan, J.M. and Chang, Chin-Yung (1970). "Nonlinear Analysis of

Stress and Strain in Soils". Proc. ASCE SMS, September.

2. Konder, R.L. and Zelasko, J.S. (1963). "A Hyperbolic Stress Strain
Formulation for Sands". Proc. 2nd Pan-American Conf. on Soil
Mech. and Founds. Brazil. Vol. 1, pp. 289-324 .
 135

PREDICTOR B

BI. PREDICTION METHODOLOGY

BI.I Sidewall Influence

Bransby and Smith (1975) report up to 14% decrease in the active

earth pressure coefficient due to sidewall friction in model tests. This
was considered negligible in wall prediction calculations and therefore
neglected.

BI.2 Calculation of Force in Reinforcement

A simple Rankine model was used to estimate lateral earth pressures

on the wall. Tributary wall areas were used to calculate reinforcement
forces in each layer. The earth pressure distribution was based on the
assumption of Ka in the lower third and Ko in the upper two thirds of
the wall height. This method was based on Reinforced Earth Company data
and personal measurements that indicate that forces in the bottom layers
of reinforced walls are always much less than predicted.

BI.3 Distribution of Reinforcement Tension

The maximum reinforcement tension was assumed to act at the Rankine

failure plane. A linear reduction of tensile force from a maximum at the
failure plane to zero at the embedded end of the reinforcement was
assumed. Constant tension from the failure plane to the front of the wall
was assumed for all layers except the top layer where the tension at the
facing was assumed to be 50% of the maximum value. This approach is based
on experience with Reinforced Earth Company walls.

B1.4 Wall Displacement

Wall deformations were calculated using reinforcement stress dis-

tributions calculated above together with a PL/AE approach. The embedded
end of the reinforcement was considered to be fixed. The modulus of the
reinforcement was taken from the initial tangent of standard Index tensile
tests curves for SR2 geogrid. Creep in the reinforcement was not consi-
dered.

BI.s Base Pressure

A Meyerhof distribution was assumed for the calculation of base

vertical pressures.

REFERENCE

1. Bransby, P.L. and Smith, I.A.A. (1975)

Side Friction in Model Retaining-Wall Experiments, Journal of
the Geotechnical Engineering Division, ASCE, GT7.
136

PREDICTOR C

Cl. METHOD OF ANALYSIS

Cl.l ~

The vertical stress distribution at the base of the wall and other
levels was evaluated by considering moment equilibrium about the centre of
the reinforced fill (refer to Figure Cl).

Although mobilised friction (<Pm') was considered to be acting at

the rear of the reinforced fill, at this stage the friction at the facing
was ignored. In performing the calculations, side friction was assumed to
reduce the effective density of the fill by about 5% and a unit weight of
fill of 17.5 kN/m 3 was used throughout.

The resulting equation obtained was as follows:

(Cl)

The stress at the front of the wall (of) now requires to be modified to
take account of friction at the facing.

Cl. 2 Step 2

Close to the facing it is assumed that friction will provide some

support to the soil and reduce the vertical stresses in this region. The
friction effect is assumed to vary linearly from a maximum at the face to
zero at a distance x (refer to Figure C2).

Considering the equilibrium conditions over a layer of thickness dz

produces the following differential equation:

do
-..J!. (C2)
dz

where the subscript w refers to conditions at the facing. Solution of the

above equation with the appropriate boundary conditions gives:

-2K tan <Pwz

2K tan <Pwq
yx
o
w
[ (1 - )e x -1 1 + yz
yx

where K lateral pressure coefficient

o vertical stress reduction at the facing

w

The vertical stresses determined in Step 1 were reduced on the basis of

the above equation.
 137

reinforced fill,
p
z
w
~
rnz

vertical stress distribution

Figure Cl Vertical Stress Distribution

C1.3 ~

To evaluate the tension distribution and the strains in the rein-

forcement, the horizontal equilibrium of an element of reinforced soil was
assessed (refer to Figure C3) and the resulting differential equation
obtained:

do
dT
VK ~ + 20 tan ~u (C3)
dx dx v

In the above equation both the earth pressure coefficient (K) and the
interface friction coefficient (tan ~il = ilw) were assumed to be func-
tions of x.
138

Figure C2 Soil/Facing Friction

dx
I- ~I

ah ah + dah
T

T+dT..- V
--- T
T

Figure C3 Horizontal Equilibrium in Soil/Reinforcement Element

Solution of the above differential equation with the appropriate

boundary conditions produced the following equation:

(C4)

where 01

0 3
 139

2/5 13 2 13.

horizontal stress at face

o vertical stress at distance x, as given in Step 1

v

v vertical spacing of reinforcement

13, [K - K m2/L2 + Kw(m 2/L2 - l)]/m(1 - miL)

a 0

(K - K - f3,m)/m 2
a w

13 3 and 13. f(a, , a 2 , 0" a., as, 13 , , 13 2 )

C1.4 Step 4

In determining the tension distribution, the horizontal stress at

the facing is required. The precise value of horizontal stress is
influenced by the mode of wall movement.

As a first step assume that the total horizontal force is given by

the Coulomb equation, i.e.

F = 1/2Kyz2 (C5)

In the above expression it is further assumed that the amount of soil

friction mobilized will be determined by the type and magnitude of the
movement. For rotation about the base (propped wall) the same friction
angle will be de~eloped at all depths. For incremental construction,
movement conforms more closely to rotation about the top. The mobilised
friction angle will thus increase with depth and may be approximated by:

~m = ~o + !,\Z (C6)

where ~o = friction an~.e at the surface

!'\ = friction angle gradient with depth

To obtain the lateral stress, the expression for total force is differen-
tiated with respect to z, i.e.

dF K + 12 yz2 ~ (K)
yz dz = °h (Cn
dZ

Now using the simple active earth pressure coefficient expression:

- sin I/l
K (C8)
+ sin 1/1

where 1/1 = ~o + !'\Z

140

Therefore the lateral stress for rotation about the top is obtained from
the following equation:

0h(top) = yz [K - ~ ( n cos W ) (1 + K)] (C9)

2 1 + sin I!J

While for rotation about the base the conventional expression is used:

°h(base) = Ka Yz (CIO)

- sin cJ>m
where K
a + sin cJ>m

In calculating the horizontal stress directly at the facing, the vertical

stress is reduced on the basis of the previous equation developed for
friction at the facing.

For both situations the maximum mobilised angle of friction was

assumed to be 30°. This friction angle was used generally throughout all
the calculations as it provided better strain compatibility with the soil.

C1.5 ~

The strain values at the required locations were calculated from

the tension distributions obtained according to Step 3 ~and isochronous
data for the reinforcement material. Over a range of 0 - 20 kN the curves
for 100 hrs and 1000 hrs are approximately linear and stiffness values
were determined accordingly.

The total displacements were calculated by integrating the tension

distribution curves over the effective length of reinforcement and
dividing the result by the appropriate stiffness.

Cl.6 ~

Taking a connection strength as 90% of the short-term strength, the

failure load was estimated by plotting load at connection versus surcharge
stress and extrapolating to determine where the curve intersected the
connection strength. It is appreciated that as failure approaches the
behaviour will be very plastic and the approach adopted is likely to be
conservative.
 141

PREDICTOR D

Dl. PHILOSOPHY OF METHOD

For the design of reinforcement layers (length, tensile forces)

limit states corresponding to internal and external stability are
assumed. For details cf. (2). Creep is prevented by restriction of the
strains corresponding to the predicted tensile forces. Empirical est i-
mates are used for predicting the displacements and strain-distributions
in the reinforcement.

D2. MODELLING OF SYSTEM

D2.1 Horizontal panel movements

Horizontal displacements at the top of the wall are assumed to be

equal to some percentage: de of the height of the wall:

de 1.5%; incremental panel wall

propped panel wall

and to vary linearly with height. The given values for de come from
observations from other large-scale tests. For the propped panel wall a
greater stiffness, with respect to the incremental panel wall, was
assumed, reflected by de = 1% and de = 1.5%. The assumed amount and
distribution of displacements is supported by experimental data cf. (2),
(3). Following the arguments of (4), creep displacements can be neglected
(experience from granular backfill and limit-analysis based design).

D2.2 Strain in reinforcement

Within our design concept, the strain-distributions f~r the rein-

forcement layers cannot be predicted. We are restricted to the following
statements:

- The strain distribution can be expected (qualitatively) as shown

in Figure D1.

c:

-...
o
I f)

distance

Figure D1 Strain in a Reinforcement Layer

142

- The maximum strain can be expected for an intermediate

reinforcement layer.

- The maximum tensile force (due to internal stability) can be

calculated as follows:

(Dl)

where y unit weight of soil

h height of the wall

llh spacing between reinforcement layers

K coefficient of earth-pressure

whereas Ka was assumed for the incremental panel wall and K

0.5 (Ka + Ko) for the propped panel wall.

Surcharge loads have been treated as follows:

(D2)

where p = surcharge load

The maximum strain (due to internal stability), therefore, corresponds to

the following tensile force:

Z = Z + Z (D3)
m p

D2.3 Total load at panel/reinforcement connection

For calculation of the connector forces, the earth-pressure

distribution from Figure D2 was assumed cf. (2).

gravity loads (y) surcharge loads (p)

pk

FigureD2 Earth-Pressure Distribution

 143

For the incremental panel wall K = Ka was assumed and, because of

the greater stiffness, K = 0.5 (Ka + Ko) for the propped panel wall.
The assumed earth pressure distribution is supported by experimental data
cf. (2).

D2.4 Fressure along base of reinforced block of soil

Due to local base failure, the vertical pressure, in particular in

the vicinity of the face of a wall, can be essentially lower than the
overburden pressure. This phenomena can be expected to be very pronounced
for a weak soil, underlying a reinforced block of soil, and to be less
pronounced for a concrete floor, underlaying a wall (RMC-test facility).
Our prediction was established for a wall resting on a soil half-space .and
is based on experiences reported by others from large-scale tests cf. (2),
(1) .

The following assumptions have been used for predicting the earth-
pressure distributions:

- At the face of the wall the lateral and vertical pressures are
assumed to be correlated via the earth-pressure coefficient (cf.
Figure D2).

The mean-vertical pressure, defined for the length of the rein-

forcement layer, is equal to the overburden pressure.

D2.S Surcharge pressure required to produce failure

Because of the great reinforcement length ~: ~ = h with ~ = 0.4 h

is usually sufficient for external stability cf. (2), no external failure
was expected. This statement is also supported by experiences from
large-scale tests reported by (3).

Back-calculation of surcharge loads due to internal stability

(failure of reinforcement or connectors) leads to a very high value:

p = 1234 kN/m 2

i. e. before a loss of internal stability can be expected, probably base

failure under the surcharge load occurs cf. (3).

D3. FARAMETERS USED IN ANALYSIS

From Bulletin No.1, Class A Frediction Exercise for (RMC) Rein-

forced Soil Walls.

REFERENCES

1. Carroll, R.G. and Richardson, G.N.: Geosynthetic reinforced

retaining walls. Froceedings of the Third International Conference
on Geotextiles, Vienna, Austria, 1986.

2. Gudehus, G. and Schwing, E.: Standsicherheit Kunststoffbewehrter

Erdbauwerke an Gelandesprlingen. Froceedings of the Baugrundtagung,
Nlirnberg, West Germany, 1986.
1M

3. Wichter, L., Risseuw, P. and Guy, G.: Grossversuch zurn Tragver-

halten einer Steilwand aus Gewebe und Mergel. Proceedings of the
Third International Conference on Geotextiles, Vienna, Austria,
1986.

4. Yarnanouchi, T., Fukuda, N. and Ikegarni, M.: Design and techniques

of steep reinforced embankments without edge supportings. Third
International Conference on Geotextiles, Vienna, Austria, 1986.
 145

PREDICTOR E

El. INTRODUCTION

This prediction utilized classical earth pressure concepts as

generally practiced in the design of reinforced earth walls using metallic
strips. The Geotechnical Eng ineering I iterature is abundant with such
design information and the performance of these walls has been excellent.
To the authors' knowledge there have been no failures nor cases of excess-
ive deformation using such design procedures.

However, even within such a design framework there are certain

design modifications and assumptions which become necessary. Our thinking
on how we arrived at these various decisions will be outlined in the text
to follow.

Lastly, it was recognized that limit equilibrium represents a

failure condition but was used to predict pre-failure situations. Thus it
was tempting to include an arbitrary reduction value on the predicted
values. This temptation was resisted on the following grounds:

(a) We could not decide on what value to use.

(b) We did not know how to reduce the factor as failure

approached.

(c) We did not know if the value would change for different parts
of the prediction request.

(d) There would be no way to back-calculate to the "correct"

answers.

E2. EARTH PRESSURES

E2.1 Options

Vertical pressure: Uniform, trapezoidal or Meyerhof (Figure El)

Horizontal pressure = K times vertical pressure
K-value: Ka , Ko or higher; constant or varying with depth
(Figure E2)
Used in final prediction: Maximum earth pressure related to
overburden+ surcharge by Ko for top
three grid levels and by Ko/2 for
lowest grid level (Figure E3).

E3. ALLOCATION OF FORCES TO GRIDS

For the incremental (Element-) wall the grids are assumed to take
on the resultant of the earth pressure per face element (O.75m times 1 m).

For the single (rigid) panel wall, the earth pressure allocated to
each grid is determined by lines drawn midway between grids and/or the
horizontal boundaries of the backfill.

These earth pressure allocations are illustrated in Figure E4.

146

14 3m
~I

I I uniform

[ , trapezoidal

I Meyerhof

Figure El Vertical Stress Distribution Options

Sketch of wall and earth pressures

Ka Ko
_--r-.-~K

z
Options:
KaKo KaKo~ K
I"I""'""--r-r-~ K rr--r-__

z z z

Figure E2 Horizontal Earth Pressure Distribution Options

 147

0.75m
--------

0.375 m
-- ~---
0 .375m

------

- (Example: Incr. Wall)

Figure E3 Earth Pressure used in Calculations

0.75m
---- - - - - -
0.375m
-- ~---
0.375m T2

- ----
TI
- per panel
(incremental wall) (rigid wall)

Figure E4 Allocation of Earth Pressure to Grids

148

E4. DISTRIBUTION OF SHEAR STRESSES ON AND TENSION IN THE GRIDS

Options considered are shown in Figure E5.

Some initial calculations were done assuming uniform distribution

of shear stresses and the resulting triangular distribution of tension.
These conditions would resemble those attained during an ideal pullout
test at the point of failure.

For the prediction, tension T was assumed to have a polynomial

distribution with Tmax at a distance of 1m (=1/3 of total length of
grid) and Tmax/2 at the face while the curvature of the distribution is
reversed in the last third of the grid to reach zero at the end. The
corresponding distribution of shear stresses is triangular, with a stress
reversal at the 1m point (Figure E6).

E4.1 Mathematically

With x distance from wall face and Tmax = Resultant of earth

pressure
0-2m: T = (0.5 + x - 0.5x 2 ) Tmax and" (I-x) Tmax
"max = (+/-) Tmax
Tavg = (11/18) Tmax = 0.61 Tmax
At face: To = Tmax/2
2-3m: T = (4.5 - 3x + 0.5x 2 ) Tmax

E5. REDISTRIBUTION OF FORCES FOR RIGID WALL

Resultants of earth pressure on each section of the rigid panel are

redistributed so that they vary linearly with height.

E5.1 Procedure

(a) Calculate resultant forces Fi (Total Force) on each part of

the panel.

.
(b) Calculate sum of all 4 forces: LFi = FR·

(c) Calculate sum of all overturning moments due to Fi:M

LFihi·

(d) Calculate parameters s and a:

Lh.
F [_1 - h 1
R 4 r
(E1)

(E2)
 149

3m

a) uniform shear b) triangular shear

I -T I +T
~
~

~
c) ideal reinforcement

.~
- T
ttr ··· IT
e) incomplete mobilization f) incomplete mobilization
(triang.) (rect. )
Figuse E5 Various Possible Distribution of Tension and Sh:ar Stresses

T.... /2l. ( ' . . IT",.,

. Tpolynomial distribution
T max = resultant of e~rth pressure per panel
Figure E6 Distribution of Shear and Tension as Assumed in Prediction
150

(e) Calculate redistributed forces as follows:

F. = a - sh. (E3)
1. 1.

(0 Calculate strains based on F avg = 0.61 F. If the modulus is

constant for all stress levels, then the deformation should
also be constant.

E6. SPECIAL CONSIDERATION OF BASE PRESSURE

A check showed that trapezoidal distribution would result in nega-

tive pressures at the back of the reinforced soil. A Meyerhof distribu-
tion was considered unsuitable for a prediction.

Chosen was an arbitrary distribution which resembles what is

generally measured below Reinforced Earth structures, although it may
still tend to be conservative as far as maximum values are concerned
(Figure E7).

o 2 3
. . . .. . . . . . . . . . . .1. . . . . . . . . . .
.
(m)

0.67

pressure
p = initial overburden stress (including 12 kPa surcharge)
q = additional surcharge

Figure E7 Assumed Distribution of Base Pressure

E7. STRAINS AND DEFORMATIONS

Initial wall deformations (incl. 12 kPa at 0 hours) were calculated

as the cumulative deformations due to earth pressure for Tavg.

Additional deformations were added for each stress increase at a

particular level (non-cumulative).

Modulus chosen was the secant modulus for the appropriate stress
level and creep time. For each stress level the appropriate curve was
selected (no correction for "accumulative" creep).
 151

PREDICTOR F

Fl. PHILOSOPHY OF METHOD

This work is aimed to present a model for the prediction of the

behaviour of a geogrid reinforced earth wall (forces and strains in the
inclusions, horizontal movements of the wall face, pressure under the
reinforced soil block). In particular, the model developed for the
present prediction refers to the Incremental Facing Reinforced Soil Test,
built at the Royal Military College, Kingston, Ontario.

The predictors have chosen to develop a rather simple method, in

order to obtain satisfactory results, carrying out the calculations with a
small computer and in a short time. The use of a simple model involved
some s impl ifying hypotheses, based only on the experience achieved from
previous field investigations on geogrid reinforced earth walls.

F2. PARAMETERS USED IN ANALYSIS

F2.1 Soil

On the basis of the results of Direct Shear Box tests and according
to the dry dens ity of the backf ill sand (1790 kg/m 3 = 17.56 kN/m 3 ) an
internal friction angle of 43° was assumed.

F2.2 Reinforcement - Inclusion

The parameters for the Reinforcement - Inclusion behaviour were

obtained from the Load - Extension data and from Isochronous Load-Strain
curves data.

The Isochronous Curves were schematize.d by means of the parabolic

equation:

£(F,t) = aCt) • F2 + bet) • F (Fl)

where the coefficients aCt) and bet) were obtained from the system:

F2 aCt)
A
(F2)
F2 bet)
B

The input data (FA,£A), (FB'£B) were obtained from the Isochronous
Curves corresponding to FA = 10 kN/m and FB = 30 kN/m.

F2.3 Soil - Inclusion Interaction

It was assumed that no relative movement between the geogrid and

soil occurred in the considered range of forces.

F2.4 Surcharge

A uniform surcharge applied according to the schedule in Table 1

was assumed.
152

F3. MODELLING OF SYSTEM

F3.1 Maximum Forces in the Inclusions

The evaluation of the forces induced in the geogrids was done by

g1v1ng a~ "influence zone" to each reinforcing layer (geogrid). The load
applied to each layer was calculated as the sum of the Rankine active
pressure and of the pressure due to the surcharge, as shown in Figure Fl.
The active pressure coefficient Ka was assumed to be constant and equal
to: tan 2 (4S0-43°/2) = 0.189.

In order to simplify the calculations, the zero reference for the

vertical axis was not positioned at the crest of the wall, but elevated by
a quantity equal to (Ws/Y). Therefore the formula used was:

(F3)

where: V. = Ws/y (see Figure Fl)

Vi = as shown on Figure Fl

The total force per unit width was calculated as:

Ttot = 1/2'Y'K '(H+W /y)2 - 1/2'y'K '(W /y)2 = E.p.(W) (F4)

a s a s 1 1 S

F3.2 Distribution of the Tensile Forces in each Inclusion

The calculation of the tensile force distribution in each geogrid

layer was based on the following assumptions:

- the tensile force versus the distance from the wall face was
represented by a curve, approximated by two straight lines,
exhibiting a maximum at a distance equal t2 xi* from the wall
face (see Figure F2);

the maximum value of the tensile force in the geogrid (Fimax)

was assumed to be equal to the above calculated Pi(W s );

- the position of Fimax was given by a line defining the "locus

of maximum tensile force"; in this case it was assumed to be at a
slope of 20°, based on the results given by YAMANOUCHI et al.
(1986) and BERG et al. (1986), as shown in Figure F2;

- the tensile force distribution in each inclusion (Fi(x) dia-

gram) was assumed to have the same slope on both sides of the
maximum: that is, the tensile load transfer by friction with the
adj acent soil was assumed to be of the same intensity on both
sides of the peak force; the remaining force at the wall face was
the load at panel/reinforcement connection.
 153

r - -
1V5 •
ZI VI Vi

- - - P4 O.9m
--- -
~P3 O.8m
-- - ~P2
H

O.7m
---
---PI O.6m
r{....'\~~A'{(1'" Vf~~,'fo.'t~

Figure Fl Scheme for the Calculation of the Forces in the Inclusions

Fj max

0.5 m I-.f

F·I

PfaC8j
'----'-=---~- ......- X

Figure F2 Tensile Force Distribution

154

Therefore the formulas used for the calculation of the tensile forces in
each inclusion and at the face were:

(Fs)

(F6)

x.*
l.
(F7)

F3.3 Strains in the Inclusions

Isochronous Load-Strain Curves were used for the calculation of

strains in the geogrids due to creep, according to the parabolic scheme
presented i~F2.2.

Following the qualitative pattern of strain versus time and sur-

charge shown in Figure F3 and taking into account the elastic strain (from
load-extension diagram at 0 h) and the creep strain (from isochronous
load-strain curves), the values of the strain in each geogrid layer at
each distance from the face and at specified elapsed times, were given by:

k-l
e: i (Ws k , x, t) {e:[Fi(Ws k , x,t), t=Ol} + ~ j ae: j (F8)
i

(F9)

k-l k-l
~ j M;. =. ~ . {e:[F.(Ws, x),at.l -
J l. J
I J
(FlO)

strain in the geogrid layer "i", due to the sur-

charge WSk

tensile force in the geogrid layer "i", due to the

surcharge WSk

ae:. increment of strain due to creep, during surcharge WS j

J

at. period of application of the surcharge WS j

J

CR(F.,at.) = In-Soil Creep Factor.

l. J

Another hypothesis (eq. F9) was that the tensile force in the geo-
grids was constant for the entire period of application of each surcharge.

The In-Soil Creep Factor takes into account the difference in creep
behaviour between laboratory and field conditions: without any useful
information on this matter, it was taken as a constant equal to 0.85.
 155

50
Ws 40
[kPa] 30 WSj

12~----""'"

100 1100 1200 1300 2200 t [hrs]

I I
I~tj I
~
[%] I I
I I
111
~E'
~_---,~~-I f "'P IJ I
f(F(~, t),'Oh)
t [hrs]
Figure F3 Qualitative Pattern of Strain versus Time and Surcharge

n
84

-X
r- I I
~4 ~ ----~-r-~1 +~2+~3
~3--y _____~~ ~1 +~2
~2~ _ _ _ _ _ ~ ~1
~1

Figure F4 Scheme for the Calculation of the Horizontal Wall Movements

156

F3.4 Horizontal movements of the wall face

To evaluate the horizontal movements of the wall face, the follow-

ing hypothesis was used: The reinforced block acts as 4 boxes superim-
posed one over the other and each of them represented by one of the
"influence zones". The horizontal pressure which increases with depth
under the crest causes the lowest box to move the other boxes with it.
The procedure is repeated for each subsequent box. The horizontal move-
ment of each box was assumed to be equal to the total elongation of each
geogrid, calculated as an integration of the strains:

L N
J ci(Ws,x,t)dx ~ ~ C •
l.
(W
S
,x., t) • LJ.
J
(Fll)
o 1j

where: ci(Ws,x,t) = strain in the geogrid layer "i"

Ai as shown on Figure F4

Assuming the above mentioned mechanism, the horizontal movement of each

"influence zone" was given by:

i N
~ • ~ Ch(Ws,xj,t) • Lj (FI2)
Ih Ij

F3.5 Pressure under the reinforced soil block

Considering that the surcharge was uniformly distributed over the

whole crest area, it was assumed that no variation of the pressure at the
base of the reinforced block versus the considered position occurred. So,
the base pressure was assumed to be a function of the surcharge Ws and
of the dry density of the backfill sand:

rex) = Ws + Y • H (F13)

F4. CONCLUSIONS

At the moment, the most important assumptions to be improved are:

the actual active pressure distribution; the position of the "locus of
maximum tensile force" line; the distribution of the tensile forces in
each inclusion, and the definition of the "In-Soil Creep Factor".

REFERENCES

1. Berg, R.R., Bonaparte, R., Anderson, R.P. and Chouery, V.E. (1986).
"Design, Construction and Performance of two Geogrid Reinforced
Soil Retaining Walls". Proceedings I I I International Confer-
ence on Geotextiles, Vienna, Austria.

2. Yamanouchi, T., Fukuda, N. and Ikegami, M. (1986). "Design and

Techniques of Steep Reinforced Embankments without Edge
Supportings". Proceedings III International Conference on
Geotextiles, Vienna, Austria.
 157

PREDICTORS G, H, I, J

These predictors did not submit prediction methodologies in the

manner requested. It should be noted, however, that predictors G, H, I
and J adopted design strategies that can be broadly classified as limiting
equilibrium approaches.

5. COMPARISON OF PREDICTED VERSUS MEASURED RESULTS

5.1 General

The types of predictions requested have been described earlier in

the paper. It should be noted that predictions were submitted for wall
performance at 100 hrs and up to 1000 hrs after each surcharge load incre-
ment was applied. As a result there is a large amount of data available
to base comparisons on. In the interests of brevity, only selected data
is compared to illustrate the success or failure of the analytical methods
employed. In general, predicted and measured test results have been
restricted to predictions corresponding to the 12 kPa surcharging incre-
ment after 100 hrs and the 50 kPa surcharge loading increment after 1000
hrs.

6. INCREMENTAL PANEL WALL

The surcharge loading schedule actually applied to the incremental

wall test is presented in the companion paper by Bathurst et al. The
schedule differs from the proposed schedule originally given to the pre-
dictors in that the initial 12 kPa surcharging increment was left on
longer than 1000 hrs and the last loading increment of 50 kPa was termi-
nated after 500 hrs (instead of 1000 hrs). Nevertheless, very little
movement in the system was recorded in the reinforced wall over the
extended 12 kPa loading increment. Similarly, after 500 hrs of the last
50 kPa loading increment, deformation rates were essentially zero and no
difference in measurements at 500 hrs or 1000 hrs would have been antici-
pated.

6.1 Horizontal Panel Deformations

Horizontal panel deformations at selected elapsed times are given

on Figure 3a and 3b. The measured data has been presented in two ways:
The curve with the larger values represents "total" horizontal deforma-
tions referenced to the initial location of the toe of the bottom facing
panel unit at the time of construction. In other words these deformations
are the sum total of deformations (or total wall out-of-alignment) incur-
red during construction and during sustained surcharging. The other curve
represents only post-construction measurements. Only predictor A consi-
dered the contribution of construction-induced deformations in his calcu-
lations. Unfortunately, his finite element model predicted values of
deformation indicating inward wall movements not observed in the model.
All other "predictors calculated deformations with respect to the end of
construction panel alignments. However, the measured data curves illus-
trate that system deformations which occurred during construction are
s ignif icant. Even at the end of the test after sustained surcharging,
fully 50% of the top panel movement was due to movements generated during
construction. The inability of analytical methods to accomodate construc-
tion-induced deformations represents an important shortcoming. Most esti-
158

inward outward

T- ____ ~ 1-.-
3.0 measurerl (post-construction deformations only)

G H _Cf' ~ _ SR2 layer 4

I, I .\' \ /, ,
~ I f measured (total deformations)
2.0 tl,' \;' /1
/1
~
5
~I/\._./
t-j,}~ - \. - -~r -- i-/o - - - SR2 layer 3

jilt
OJ II/
\//\
/ ..
I I
II
I.od' ~- - -1- ~~- ~- --- - -- SR2layer 2
:~!
: Il /
/ .'",{/
r /// \12 kPa surcharge for 100 hrs \

o
i "///~:'f/
-0- - -.0 - / r - - - - - - - - - -- SR2 layer 1

horizontal panel movements (mm)

a) Deformations after 12 kPa surcharge for 100 hrs

1--1---1-__ ~~I~e~~1-~ _T-_tl96

inward outward
(post-construction deformations only)
3.0 measured (@ 50 kPa surcharge for 500 hrs)

,
G_
\ / \ (total deformations) 1
, . . measured (@ 50 kPa surcharge for 500 hrs) I
'\'I / \. . . . . '" I
L W- -i - - ~~2~a~r] i 9<:- ---d- - - - -1- -
2.0' //
I _1!l9
.§ I I \. 1 I ....( / if
] /.1I' . . '" I
",. /
/ I
I
.,
~ 1.0
I
- aI-. ---/-
'" /
--- ~. .....--.
.....
. .-..t - ~R': 1~'2:
....
:'..-J:> - - -
",-_""_______-, .-_~
_109

\\ / '" . . . --- 1. ....150 kPa surcharge for 1000 hrs

I

I
vt . . ___- -. . . -. . . - - - <>. . /--- . . .
0;-/ ,,/ ,,"""
I SR2 layer 1
6 ~ -/1- - . - - - - - - - - - - - - -

horizontal panel movements (mm)

b) Deformations after 50 kPa surcharge for 1000 hrs

Figure 3 Panel Deformations (Incremental Wall Test)

 159

mates of deformations were significantly higher than measured values des-

pite neglecting construction deformations. The limiting equilibrium based
methods employed by predictors G and H for post-construction deformations
generated values which can be considered in the correct range.

6.2 Strain in Reinforcement

The strain in the geogrid reinforcement at selected times is given

on Figure 4a and 4b. Similar to the comments made concerning horizontal
panel movements, the measured strains include strain accumulated as a
result of construction activities. Again only predictor A considered con-
struction-induced strains in his prediction. All others referenced strain
levels to the end of construction condition. All measured strains were
less than 1% and the companion paper by Bathurst et al. notes that in
layer 1, for instance, more than 50% of the recorded strain at the end of
the test (i. e. 50 kPa surcharge for 500 hrs) occurred during construc-
tion. With the exception of predictor A, all analytical methods pre~icted
values of strain which are in excess of the strains which could be associ-
ated with post-construction surcharging activities. The limiting equili-
brium based methods considered to be least in error can be ascribed to
predictors G and H.

6.3 Load in Panel/Grid Connections

Predicted and measured connection loads at the end of the test are
given on Figure 5. Of all the predictions offered, only the finite-
element model employed by predictor A gave values which were consistent
with the measured values. Other predictions typically over-predicted the
test results.

6.4 Vertical Pressure Along Base of Reinforced Soil

Predicted vertical pressures at the end of the test are given on

Figure 6. No reliable physical measurements of vertical earth pressures
were obtained from this test. It is interesting to note that predictions
A,C,D and E gave similar trends in estimated values. In addition, these
predictions can be seen to straddle estimates based on the simplified
assumption of uniform vertical earth pressure according to yh+q.

6.5 Predicted Failure Surcharge Loads

The RMC test facility at the time of the workshop was only capable
of generating surcharge loadings up to 50 kPa. It was recognized by the

Predictor Estimated Surcharge Load

to Initiate Failure (kPa)

A +200
B 350
C 130
D 1234
E 110
F 365

Table 2 Predicted Values of Surcharge Load to Initiate Failure

of the Incremental Panel Wall Trial
160

Predictor Symbol
A 0
B Cl
SR2 layer 4 C 0
D \J

••
2 E 0
F
0 0 0
•
0 G
H

0
2 0 0 0 0
SR2 layer 3
,..--..,
~
'--'
...., Cl
e 0 !!!l 8
>=:
v i
S
v
0
u
J-< 0
.8 0 0 0 0
.....>=:v 2
J-<
SR2 layer 2
0 0 0
.....
>=:
H 0
•
Cl
.....>=:cd
....,J-<
rJJ
• II
0
2 SR2 layer 1
Cl Cl
112 kPa surcharge for 100 hrs I

distance from back of panel (m)

Figure 4a Strain in Reinforcement after 12 kPa Surcharge for 100 hrs
(Incremental Wall Test)
 161

Predictor Symbol
SR2 layer 4
0 A 0
0 0 B t::.
0 C 0
t::. D 'il

••
4 t::. E 0
t::. F
t::. t::.
•
0 G
0 0
• •
H

• ••
0
2
• •
0 measured
0
0
0 (50 kPa @ 500 hrs)
0
0 0 0 0
,.....-.,
4 SR2 layer 3
8 c
~ t::. t::.

• •
'---"
+" 0
8
~
Q) 0
S 2 measured

.8
Q)
u
I-< • (50 kPa @ 500 hrs)

......
~
Q)
0
I-<

4 0 0 @ ~
•
......
~ 0 SR2 layer 2
......
~
cd
I-<
i • • 0
t::. t::. t::. et::.
• • •
+"
'n 2
measured (50 kPa @ 500 hrs)

0
4

~
• •
t::.
50 kPa surcharge for 1000 hrs
2 0 8 measured (50 kPa @ 500 hrs)

00 0.5
distance from back of panel (m)

Figure 4b Strain in Reinforcement after 50 kPa Surcharge for 1000 hrs

(Incremental Wall Test)
162

r 1 rf-",
3.0

~~ ;,,' range
', of measured
\ \ \ I
U; ~{pa 500 h\ \1
/ values \
2 ,0
@

:§: I "
~
,9
... I \\ \ /
~
I ' , \
\
~
-.; /
1.0
~ \ \ /p 'rJ _0

c?'t \- -
I \
I '
\
'/
'y _r-- \
I / \
t&---; 0- j \
00 2 4 6 8 10 12 14 16 18 20

connection load (kN 1m)

Figure 5 Connection Loads (Incremental Wall Tests)

130

120
cd
0...
~
....
Q)
110
;l
en
en
Q)
....
0..
100
Cdu
'~
....
Q)
> 90
50 kPa @ 1000 hrs
no measured data

80

10 0

distance from back of panel wall (m)

Figure 6 Vertical Pressure along Base of Reinforced Soil

(Incr e mental Wall Test)
 163

researchers at RMC (and by the predictors) that this load level would be
insufficient to fail the trial walls. Nevertheless, the predictors were
invited to submit their estimates of the critical surcharge loading
required to initiate failure. The predicted values fall over a wide range
as shown on Table 2.

7. PROPPED PANEL WALL

The number of predictions offered for aspects of the propped wall

behaviour was significantly less than the response for the incremental
wall. This observation was a surprise to the workshop organizers because
the propped wall behaviour is a relatively easier problem since the facing
unit was restricted from moving until after the backfill had been placed
to the full wall height.

7.1 Horizontal Panel Movement

The predicted and measured wall movements at selected times are

given on Figure 7a and 7b. All methods resulted in deformations which
were in excess of those measured. Nevertheless, all predicted near-rigid
rotation of the facing panel about the toe. Predictor A gave the smallest
estimates for predicted wall deformations and hence closest to those
actually observed.

7.2 Strain in Reinforcement

The strain in the geogrid reinforcement at selected times is given

on Figures 8a and 8b. With the exception of predictor A, all participants
significantly over-estimated measured strains in the reinforcement which
never exceeded 1%. However, no analytical method could account for the
observed increased strain gradient in the vicinity of the top reinforce-
ment connection considered to be the result of relative downward movement
of the sand backfill behind the wall.

7.3 Load in Panel/Geogrid Connections

Predicted and estimated values of panel/geogrid connection loads

are given on Figure 9.

Predictors A,I and J estimated low to non-existent connection loads

for the propped panel wall at the end of the test. The measured values
are also low and comparison of estimates shows that predictor A using a
finite element model and predictors I and J using limiting equilibrium
methods were "qualitatively" correct.

7.4 Vertical Pressures at the Base of the Reinforced Soil

Predicted and measured values of vertical earth pressure are plot-

ted on Figure 10 for the end of test condition. Essentially all predic-
tions and ' the measured values of vertical earth pressure showed the same
trends with relatively high earth pressures recorded 1 m back from the
wall and reduced values toward the wall and below the free end of the
reinforced block of soil. The discrepancy between measured and predicted
values is likely due to loss in base pressure due to sidewall friction and
(possibly) some under-registration of cell response due to cell/soil
interaction.
164

rr
outward

1
3.0

measured II /
/
2.0
I II /
I II /
:§: ? I
'<70
1 ?
112 kPa surcharge for 100 hrs I
/
"
.S
10...
I
/
I I /
I / I /
"
?1 /
<l
1.0 d I
0
/
/ / / I
// I I
/i rf
/

20 30 40 60 70 80 90 100
horizontal panel movements (mm)

a) Deformations after 12 kPa surcharge for 100 hrs

outward

r
3.0 measured

f /
I
T
I
I /
/
~103
2.0 / /
I /
:§:
"
.S
/' /
J> ,P
/ '
10... / / /
-ll / / /
/ / /
1.0 ? "P ./
"P
I I ./
./
I ./ ./
I I ./
./
./
./
/1 ./
0 if 50 kPa surcharge for 1000 hrs

horizontal panel movements (mm)

b) Deformations after 50 kPa surcharge for 1000 hrs

Figure 7 Panel Deformations (Propped Wall Test)

 165

Predictor Symbol
4 A o SR2 layer 4
C o
o
o•
E
I
J measured
2

0
4
SR2 layer 3
..--.,
~
2
+'
~
Q)

S
Q)
U 0
..s
I-<

4
......
~
Q)
SR2 layer 2
I-< measured
......
~

......
~ 2 0
C'd
I-<
+'
rn
8 0.8 0

0
4
SR2 layer 1
measured
2
112 kPa surcharge for 100 hrs )

00 0.5 1.0
distance from back of panel (m)

Figure 8a Strain in Reinforcement after 12 kPa Surcharge for 100 hrs

(Propped Wall Test)
166

Predictor Symbol
A 0
C 0

•
E 0
4 I
0 0 0 J
0 °
0 0 ° @
2 0
0

0 • 0
0
4 0 0 0 0 SR2 layer 3
------
~ 0
0
8
8
•
'--" 0
+" 2 0
>::
Q)

S
Q)
U
I-<
0
...s>::
..... 4 0 0
Q)
I-<
0 SR2 layer 2
0
.....>:: §
.....>::
cd
I-<
2
0°0
• 0
0
+"
[{J

0
4
SR2 layer 1

~
2
8 50 kPa surcharge for 1000 hrs

00 0.5 1.0 1.5 2.0 2.5 3.0

distance from back of panel (m)
Figure 8b Strain in Reinforcement after 50 kPa Surcharge for 1000 hrs
(Propped Wall Test)
 1
167

m"," ,,' ".'000'"' 00'''.,

'} 1 ! f--",
I

/ / "- I
2.0 / I "'- I
"'-
I
~
0
~
1I t---I

/
f ~
\
~23.8
I
1
~ \
I
~H
fI I
50 kPa surcharge for 1000 hrs I
I
~ 27.6

-- -- -- --
___ ___ 1 213

\ I
\I
~
I
~---
I
-- - 1
1
I
~23.5

00 2 4 6 8 10 12 14 16 18 20
load in connections (kN / m)

Figure 9 Connection Loads (Propped Wall Test)

130
50 kPa surcharge for 1000 hrs

120
"?
P-.
C
~ 110
;::I
[fJ
[fJ
Q)
.....
0..
100
~
.~
...-
.....
Q)
:> 90
measured
80

70

600~----e-----~----------~~--------~
2.0 3.0
distance from back of panel (m)
Figure 10 Vertical Pressure along Base of Reinforced Soil
(Propped Wall Test)
168

7.5 Surcharge Loading to Initiate Wall Failure

The estimated magnitude of surcharge loading to initiate failure is

given on Table 3.

Predictor Estimated Surcharge Load

to Initiate Failure (kPa)

A +200
C 101
E 110
F 365
I 540
J 532

Table 3
Predicted Values of Surcharge Load to Initiate
Failure of the Propped Panel Wall Trial

Similar to the results of the incremental panel wall predictions,

there are a wide range of estimates of the critical failure load to
initiate failure.

8. DISCUSSION

At the outset it should be noted that the predictors were faced

with a formidable task: Specifically, they were required to model the
behaviour of a sand/polymeric material composite. The constituent parts
that comprise this composite are themselves difficult materials to model.
Furthermore, the synergism between components (interaction, arching, etc.)
has yet to be formulated.

With the exception of predictor A the methodologies employed were

essentially limiting equilibrium analyses. The use of limiting
equilibrium models to predict the behaviour of the tri~l walls which were
at working load conditions is questionable at best and possibly very
inaccurate. Nevertheless, the predictors recognized this general
deficiency and in most instances attempted to modify their methodologies
to accomodate this contradict ion. For example: deformations far less
than those required for plastic conditions in soil; friction angles below
peak values; and earth pressure coefficients greater than Ka.

While most predictors gave geogrid strains that were significantly

higher than measured, all strains were less than the 10% performance limit
for the inclusions (McGown et al. 1984) confirming that analytical
proce?ures correctly predicted that rupture of the reinforcement would not
occur under the surcharg~s and surcharge durations applied. It can be
inferred from the strain data predictions that most analyses predicted not
only larger strains but propagation of these elevated strains further from
the wall facings than was actually observed.
 169

Based on the Statements of Methodology supplied by the participants

as part of the Class A predictions, none of the limiting equilibrium
methods were able to differentiate between incremental and propped wall
construction and facing type. This is a serious shortcoming since signi-
ficantly different facing deformation behaviours were recorded for the two
retaining wall systems. In particular, the majority of facing movements
for the incremental wall occurred during construction.

Limit equilibrium methods may be sensitive to the choice of fric-

tion angle employed particularly as the soil was likely at a pre-failure
condition. The choice of friction angle is sensitive to the interpreta-
tion of direct shear data and further complicated by the (near) plane-
strain conditions of the RMC Test Facility. The peak shear friction
angles used in limiting equilibrium methods should be increased to account
for plane-strain conditions as pointed out by Jewell (1987a, 19117b),
Bonaparte and Schmertmann (1987). The unreinforced test reported by
Bathurst and Benj amin (1987) in these proceedings confirms that a rela-
tively high soil friction angle may be operative in retaining wall models
when these structures are surcharged to failure.

Another shortcoming of all analytical methods was the neglect of

fac ing/ sand interaction. The results of the unreinforced propped wall
test showed that a facing/sand friction angle of 33 degrees was mobi-
lized. Consequently, a significant portion of vertical soil stress may be
transferred to the facing in the reinforced trial walls (although these
boundary forces were not monitored). This transfer should be accounted
for in calculations that consider (global) reinforced soil block stabili-
ty.

Facing/sand/reinforcement behaviour was also influenced by the

relative vertical movement between the facings and sand backfill in these
trial walls. It can also be appreciated that the degree of compaction and
support offered to the geogrid in the vicinity of the connection can have
an important influence on the strains generated in the extensible rein-
forcement. These effects are largely related to quality control during
construction and consequently difficult to implement within analytical
models.

Only two of the predictors considered the influence of sidewall

friction directly within calculations. An explanation for this may be
that only an estimate of sidewall friction angle was provided. The paper
by Bathurst and Benj amin shows that about 14% of the total active earth
force may be resisted by sidewall friction. Nevertheless, assuming that
this value is representative of the reinforced trial walls, this reduction
is not likely to bring the majority of calculated connection loads to
values close to measured results. Unfortunately, it is difficult at this
point in time to quantify the contribution of sidewall friction to the
generation of strain in the reinforcement inclusions. Bonaparte and
Schmertmann (1987) suggest that the frictional boundary conditions of the
RMC test facility may reduce reinforcement strains by 20 to 30%. Jewell
(1987b) suggests reductions up to 50%. Nevertheless, for most of the
predictions, even appl ication of these large correct ions does not bring
calculated values into the range of measured values.

In general, 1 imi ting equil ibrium methods over-est imated connect ion
loads, facing displacements and strains in the inclusions. However, from
170

an engineering point of view the methods are inherently safe. Perhaps the
most important criticism of these methods is that they are excessively
safe and hence impose an economic penalty when extensible reinforcement is
used in these structures. The finite element approach adopted by predic-
tor A generally gave the most accurate predict ions. Nevertheless, the
model did fail dramatically at some stages in trial wall testing. Speci-
f ically, the model predicted inward facing movements at low surcharge
levels. This observation suggests that despite giving relatively good
estimates of wall performance at the end of each trial wall, much work
remains to be done before FE models (and their "drivers") can be used to
predict the full range of behaviour for these structures.

Nevertheless, it appears safe to say that FEM approaches hold great

promise as the analytical tool of choice for polymeric reinforced retain-
ing walls of the type constructed for the prediction exercise. FEM are
particularly attractive if predictions of wall performance at a non-
failure state are the goal. From the practitioners perspective, however,
there are strong arguments for the continued use of limit equilibrium
methods. Certainly, i f avoidance of catastrophic failure is the only
concern then the errors associated with this class of solutions may be
justified. Such a justification is warranted provided it is not "excess-
ive" (that term varying from case to case). It is felt that this set of
predictions has focused on the magnitude of errors. Limiting equilibrium
methods cannot be a rational design approach if "factors of safety" are
simply compounded. These approaches must be used with constraint and
knowledge. It is hoped that further analysis of the wealth of information
found in these proceedings may shed additional light in this direction.

Finally, the authors would like to direct the readers to other

papers by symposium participants that employ limit equilibrium-based
approaches to solve the general problem. These include Jewell (1987a,b),
Bonaparte and Schmertmann (1987), Gourc. et al. (1987) and Delmas et al.
(1986) .

ACKNOWLEDGEMENTS

The authors would like to thank all those Workshop participants who
gave predictions of the performance of the RMC trial walls. In addition,
the authors would like to express their appreciation to all the partici-
pants for their contribution to the discussions that were held during the
Workshop particularly with regard to interpretation of trial wall behav-
iour and results of analytical models. Finally the authors would like to
thank J. DiP ietrantonio who drafted the figures for this paper and Miss
Martina Lahaie who typed the manuscript.

REFERENCES

1. Bathurst, R.J. and Jarrett, P.M. (1986). Class A Prediction Exercise

for Reinforced Earth Walls, Bulletin No. 1 for NATO Advanced Research
Workshop, Application of Polymeric Reinforcement in Soil Retaining
Structures, Depts. of Civil Engineering, RMC and University of
Strathclyde.

2. Bathurst, R.J., Wawrychuk, W.F. and Jarrett, P.M. (1987). Laboratory

Investigation of Two Large-Scale Geogrid Reinforced Soil Walls,
Application of Polymeric Reinforcement in Soil Retaining Structures,
 171

NATO Advanced Research Workshop, Royal Military College of Canada,

June 1987.

3. Bathurst, R.J. and Benjamin, D.J. (1987). Preliminary Assessment of

Sidewall Friction on Large-Scale Wall Models in the RMC, Test
Facility, Application of Polymeric Reinforcement in Soil Retaining
Structures, NATO Advanced Research Workshop, Royal Military College
of Canada, June 1987.

4. Bonaparte, R. and Schmertmann, G.R. (1987). Reinforcement

Extensibility in Reinforced Soil Wall Design, Application of
Polymeric Reinforcement in Soil Retaining Structures, NATO Advanced
Research Workshop, Royal Military College of Canada, June 1987.

5. Delmas, P., Berche, J.C. and Gourc, J.P. (1986). Le Dimensionnement

des Ouvrages Renforces par Geotextile: Programme Cartage", Bulletin
de liaison des Laboratoires des Ponts et Chaussees, 142 mars, avril
1986.

6. Gourc, J.P., Ratel, A. and Gotteland, Ph. (1987). Design of

Reinforced Soil Retaining Walls: Analysis and Comparison of Existing
Design Methods and Proposal for a New Approach, Application of
Polymeric Reinforcement in Soil Retaining Structures, NATO Advanced
Research Workshop, Royal Military College of Canada, June 1987.

7. Jewell, R.A. (l987a). Reinforced Soil Wall Analysis and Behaviour,

Application of Polymeric Reinforcement in Soil Retaining Structures,
NATO Advanced Research Workshop, Royal Military College of Canada,
June 1987.

8. Jewell, R.A. (1987b). Analysis and Predicted Behaviour for the Royal
Military College Trial Wall, Application of Polymeric Reinforcement
in Soil Retaining Structures, NATO Advanced Research Workshop, Royal
Military College of Canada, June 1987.
AN ASSESSMENT OF THE RESISTANCE OF TENSAR SR2 TO PHYSICAL
DAMAGE DURING THE CONSTRUCTION AND TESTING OF A REINFORCED
SOIL WALL

D I BUSH - NETLON LIMITED, ENGLAND

D B G SWAN - UNIVERSITY OF STRATHCLYDE, SCOTLAND

1. INTRODUCTION
Damage during construction procedures is one of the
factors affecting the physical properties of geotextiles
and related products used to reinforce soil structures.

This paper makes an assessment of the damage to and

change in physical properties of the 'Tensar' SR2 geogrids
used as reinforcing elements in an incremental panel wall
constucted at the Royal Military College, Kingston,
Ontario.

Prior to shipment the "as produced" reinforcement was

sampled. Constant rate of strain tensile and sustained load
tests (creep) tensile tests were carried out on those samples.
At the end of the experimental programme recovery of the
reinforcing elements was supervised. A visual assessment of
the damage to the "recovered" reinforcement was carried out
and the tensile tests were repeated on representative samples.

2. CONSTRUCTION, SURCHARGING AND DISMANTLING OF THE

REINFORCED SOIL WALL
'Tensar' SR2 reinforcing elements were installed and
tensioned to remove the slack. An angular gravelly sand,
with maximum particle size of 6mm, was placed by
mechanical shovel. The fill was compacted into l25mm
thick layers with six passes of a vibratory plate
compactor in accordance with the Department of Transport
Specification for Road and Bridge Works (1)

The reinforced soil wall was 3 metres high and

surcharged with an overburden pressure of 50kPa for a
period of 500 hours.

The overburden was removed from the wall with shovels

and each level of reinforcement was carefully exposed for
inspection, recovery and testing.

173

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 173-180.
© 1988 by Kluwer Academic Publishers.
174

3. TEST METHODS AND RESULTS

3.1. Sampling
Samples of the "as produced" geogrid were taken at
random from the four rolls of Tensar SR2 prior to
shipment. One roll was used to construct the "incremental"
panel wall and the "recovered" specimens were all from this
roll.

3.2. Constant rate of strain tensile tests

Samples of 'Tensar' SR2 geogrid approximately 480mm long
x 400mm wide were held in specially designed jaws,
described by McGown et al (2). They were subjected to a
constant rate of strain of 2% per minute in a standard INSTRON
1107 tensile test machine. Load and extension between the
jaws were recorded on an automatic chart recorder. Load/strain
graphs are shown in Fig.l.
All tests were carried out at 20° + 2°C.

80

70
------
60

E
........ 50
Z
~

Cl 40
«
0- I
AS PRODUCED
30 (average from 8 tests)
--- - - RECOVERED •
(average from 16 tests)
20
Strain rate: 2 %/minute
Temperature: 20o±2°C
10 Specimen size: 5 bars long by 15 ribs wide

5 10 15 20 25 30
STRAIN (%)
Figure 1
AVERAGE LOAD·STRAIN CURVES FROM CONSTANT
RATE OF STRAIN TENSILE TESTS ON TENSAR SR2
 175

Th e r es ults fo r " as p r od uced" a nd "recove r e d"

re in fo r ceme nt are s h own in t ab l es 1 a n d 2 r espec ti ve l y a nd
s umma ri sed i n t ab l e 3 . Value s of loa d and st ra i n ar e
recorde d a t pea k loa d and ruptu re . Sti f fn ess ( i . e . s e can t
v al ue ) is r e po rte d a t 5% an d 10 % s tra in .

Peak Load Data Stiffness Values (kN/m) Rupture Data

Specimen at Various Strains
Number
Load (kN/m) Strain (%) 5% 10% Load (kN/m) Strain (%)

1 67.17 14.5 691 584 65.62 24.1

2 67.47 14.7 698 588 65.72 23.8
3 66.00 15.7 680 576 64.35 25 .2
4 66.29 15.1 678 575 64.32 26.0
5 64.83 16.6 653 555 63.30 27.0
6 65.12 16.1 656 557 63. 26 26.8
7 66.29 15.1 675 574 64.87 2 5.6
8 66.00 15.7 677 5 75 64.53 24.8

Table 1: Data from Constant Rate of Strain Testing of As Produced Tensor SR2 Geogrid
used in 'Incremental Panel' Wall.

Peak Load Data Stiffness Values at Rupture Data

Specimen Position in Wall Various Strains (kN/m)
Number Load Load
Strain(%) Strain (%)
(kN/m) 5% 10% (kN/m)

1 P.1 (Right); 65.85 16.85 710 567 65.12 19.10

B.17-21 R.12- 28
2 P.1 (Right); 65.85 16.45 698 567 64.24 20.55
B.22-26 R.1 2-28
3 P.1 (Left); 65.56 15.95 698 564 63.95 19.20
B.3-7R.1-1 7
4 P.1 (Left); 65.70 13.95 719 578 As Peak As Peak
B.8-12 R.1-1 7 Laad Load
5 P.2 (Left); 66.73 15.80 686 565 65.12 19.80
B.8-12 R.1-17
6

7
P.2 (Left);
B. 3-7 R.1-17
P.2 (Right);
B.24-28 R.12-2B
55.73

66.59
10.05

17.35
692

692
557

563
As Peak
Load
65.41
. As Peak
Load
20.90

8 P.2 (Right); 65.41 16.20 686 575 63.65 20.90

B.19-23 R. 12-28
9 P.3 (Right); 65.18 14.65 698 569 As Peak As Peak
B.14-18R.1-1 7 Laad Laad
10 P.3 (Right); 65.41 15.90 704 565 As Peak As Peak
B.19-23 R.12-28 Load Load
11 . P.3 (Left); 66.00 16.15 702 558 As Peak As Peak
B.3-9 R.12-28 Load Laad
12 P.3 (Left); 64.97 16.20 681 558 64.53 18.20
B.8-12 R.1-17
13 P.4 (Right); 65.12 16.10 704 570 63.95 20.15
B.13- 17R.12-28
14 P.4 (Right); 66.59 15.95 669 563 64.53 21.95
B.23-27 R.12-28
15 P.4 (Left); 61.89 12.80 692 563 As Peak As Peak
B.3-7 R.1-17 Laad Laad
16 P.4 (Left); 65.27 16.45 681 557 63.80 21 .10
B.8-12 R.1-17

Table 2: Data from Constant Rate of Strain Testing of Damaged Tensor SR2 KeY' P. = Ponel Number
Geogrid used in ' Incremental Panel' Retaining Wall B. = Bors
R. =Ribs
176

Peak load
Number 0/
Specimens Standard Coefficient 0/ Strain at Rupture Strain at
Mean, X' Deviation, Variation an -1 Peak load load Rupture
(kN/m) an-l (kN/m) (%) X (%) (kN/m) (%)

"As Produced" 8 66.15 0.90 1.4 16.2 64.50 25.4

"Recovered" 16 64.87 2.67 4.1 15.4 64.01 20.2

5 % Stiffness 10% Stiffness

Coefficient 0/ Coefficient of
Mean,X Standard Mean,X Standard
Deviation, Variation Variation
(kN/m) (kN/m) Deviation
a,-l (kN/m) a,-l (%) a -1 (%)
X
a n -1 (kN/m) n X

"As Produced" 676 15.4 2.3 573 11.6 2.0

"Recovered" 694 12.4 1.8 565 6.1 1.1

Table 3: Con~tant Rate 0/ Strain Data from Tensor SR2 Geogrid used in 'Incremental Panel' Retaining Wall.

3.3. Sustained load (creep) tensile tests

To determine the load-strain-time properties of
'Tensar' SR2 geogrids, sustained load tensile tests were
carried out (McGown et al (2)). The load was applied
smoothly within 5 seconds to a specimen approximately
600mm long by 400mm wide and it remained constant for the
duration of the test. Extension measurements were taken
of the specimen at pre-defined time intervals varying
from 0.2 seconds during the first ten seconds of the test
to once per day after the first 24 hours. All tests were
carried out at 20 0 + laC.

Four loaded levels were used to compare the "as

produced" geogrid physical properties with "recovered"
geogrids. The strain-time graphs are shown in Fig.2 for
each load level. The variation of rate of strain with
strain is shown in Fig.3. Isochronous curves are shown
in Fig.4.

3.4. Visual assessment of construction damage

Standard terms are used to express the physical damage
sustained during construction by each geogrid specimen.
Thes~ are:
3.4.1. GENERAL ABRASION is a subjective description of
the condition of the surface of the test specimen. The
amount of abrasion is the extent of the abraded area
expressed as a percentage of the whole specimen area.
 177

1week 1 month

---
--- --- 28.7

------ ASPRODUCED
- - - - - RECOVERED
~----------4----------2+------------+---4-------~4-------~--

-2 o 2 3
TEST TEMPERATURE 20 o ± 1°C LOG 10 TIME (hrs) LOADS EXPRESSED
Figure 2 IN kN/m width
COMPARISON OF LOAD-STRAIN-TIME CURVES FOR 'TENSAR' SR2

STRAIN(%) 45
1 hour I. 1
o 5 10 15 20 I I /10hours1
o \
f"'
~;' 100 hours

\
f

r-r-
\ ' 40
\ \ ~ /~ ~

I r1J'
I
\ \ \ / I
1\\\\,\
-1 /1000 hours
~\\\\1
/1
35

i/l, ~

' '-
\

I~1'\\
j ;,; I
30 /;j/
/ I
.I
/Ir VIvi
\ I\ --42.5

\ \\
z I
~ 1--39.6
W
~ 25
o I
\
\ 9 /, j,

!/;
\ 20
\ \ 35.8 'I

r
1 \
-5 f--- 35.1 ...
15
---t-- AS PRODUCED

,/I
28.7
I

-6 r--_
- __ AS PRODUCED ~ -1--'- REC?VERE?
- - - - - - RECOVERED 10
LOADS EXPRESSED IN kN/m width TESTTEMPERATURE: 20'± 1'C
-7
TEST TEMPERATURE: 20 D± 1 DC 5
Figure 3
COMPARISON OF "SHERBY-DORN"
CURVES FOR TENSAR SR2 o
o 2 4 6 8 10 12 14 16
STRAIN (%)
Figure 4
COMPARISON OF ISOCHRONOUS LOAD-STRAIN
CURVES FOR TENSAR SR2
178

3.4.2. a SPLIT is identified by the passage of light

through the polymer structure when the specimen is held
up against a white background. It can occur in the
longitudinal ribs or the transverse bar.
3.4.3. a BRUISE is a splayed or flattened area of a bar
or rib.
3.4.3. a CUT is the total severence of a rib.
3.4.4. ~FIBRILLATION is caused by a sharp object
moving between or across the ribs, shaving away the rib
edges.

The amount of damage sustained by each of the

"recovered" test specimens and their position in the wall
is recorded in Table 4.

Panel 1 is the lowest level of geogrid and Panel 4 the

highest level above the wall base.

Number of Generol
Specime~
Position in Wall
Number Splits Bruises Cuts ~dge Abrasion
Fibri"ation ('Yo ofTotal Area)

1 P.1 (Right); B.17-21 R.12-28 0 0 0 0 20

2 P.1 (Right); B.22-26 R.12-28 0 0 0 0 20
~ 3 P.1 (left); B.3-7 R.1-17 0 0 0 0 20
Q)
E 4 P.1 (left); B.8-12 R.1-17 0 0 0 0 20
'uQ)
a. 5 P.2 (left); B.8-12 R.1-17 0 0 0 0 10
0
'" c:
6
7
P.2 (left); B.3-7 R.1-17
P.2 (Right); B.24-28 R.12-28
0
0 0
0
0
6
0
20
10
~ 8
9
P.2 (Right); B.19-23 R.12-28 0 0
1
0 0 10
0 P.3 (Right); B.14-18 R.1-17 0 0 0 10
-5 10
11
P.3 (Right); B.19-23 R.12-28 0 0 0 0 10
""1:0 12
P.3 (left); B.3-9 R.12-28
P.3 (left); B.8-12 R.1-17
1
0
0
0
0
0
0
0
20
10
1;;
c:
13 PA (Right); B.13-17 R.12-28 0 0 0 0 10
u
0 14 PA (Right); B.23-27 R.12-28 0 0 0 0 10
15 PA (left); B.3-7 R.1-17 0 0 0 6 10
16 PA (left); B.8-12 R.1-17 0 0 0 0 10

Q) ~
"Q)c:
Q) ~ Q)
a.
1
2
P.1 (Centre); B.24-29 R.1-17
P.2 (Centre); B.23-28 R.28-44
0
0
0
0
0
0
. 0
0
10
10
'6~'5
1;;oQ)
3 P.3 (Centre); B.24-29 R.2B-44 0 0 0 0 10
:>00. 4 PA (Centre); B.21-29 R.1-17 0 0 0 0 10
"' ..... '"
Table 4: AmJunt af Visually Assessed Damage from Selected Tensor SR2 Test Specimens.

4. DISCUSSION
The size of the test specimen was chosen so that the
behaviour of the nominal one metre wide 'Tensar' SR2
reinforcing element would be replicated and the influence
of the construction damage would be properly represented.
Extension measurements during creep tests were taken on
the specimen to eliminate possible jaw effects.
 179

The geogrids suffered very little damage due to

construction, surcharging and recovery. There was only
one split and one bruised rib in the 1,500 ribs inspected.
The surface of the specimens suffered between 10% and 20%
general abrasion which was to be expected for granular
fills of 6mm maximum particle size. Greatest levels of
abrasion occurred on the lowest levels of reinforcement.
The edge fibrillation recorded on 6 ribs of each of two
specimens is thought to have occurred during recovery of
the reinforcement and to have been caused by a shovel.

In the constant rate of strain tests the "as produced"

specimens show a slightly higher average peak load and
average strain at peak load than the "recovered"
specimens. Both "recovered" specimens with edge
fibrillation damage show significantly reduced InaG and
strain at peak. The "recovered" specimens with a bruise
and a split show no significant difference, a result
confirmed by those specimen& suffering only general
abrasion. Thus the reduction of the average peak load and
corresponding strain of the "recovered" specimens is,
therefore, due to "edge fibrillation". The values of
stiffness at 5% and 10% strain show no significant
difference between the "as produced" and all "recovered"
specimens. The "recovered" specimens show a lower
variation in stiffness and this may be due to them being
selected from the same roll of 'Tensar' SR2 rather than1
taken at random from four rolls as the "as produced"
specimens were.

In the sustained load (creep) tensile tests there is

close agreement between the strain-time performance of the
"as produced" and "recovered" specimens. Examination of
the rate of strain - strain curves shows the instability
limit strain of 'Tensar' SR2 to be approximately 18% for
"as produced" and "recovered" specimens. The isochronous
stiffness curves confirm the close agreement in the strain-
time performance of both specimens.

In the analysis of reinforced soil structures the

value long-term strength, for the appropriate design
conditions, has a partial factor of safety applied to take
account of variations occurring within the production
process and the influence of construction the value of this
partial factor of safety was recommended by Netlon Limited(3)
as in the range of 1.25 to 1.4. This value was extrapolated
from full-scale trials using well-graded fills of I'1aXiI'111f'.
particle size in the range 10mm to 125mm and appears to b:?
conservative.

The properties of the "as produced" 'Tensar' SR2

geogrids were given to participants for predictions.
These can be used to represent the in service geogrids.
180

5. CONCLUSIONS
This project offered the opportunity to examine 'Tensar'
SR2 geogrids used as reinforcing elements of a soil
retaining wall constructed under well defined conditions.
Geogrids recovered from this wall suffered very little
damage in the gravelly sand fill due to the
construction,.testing and recovery processes. There was
no significant change in their physical properties
derived from constant rate of strain tensile testing and
from long-term sustained load tests. The properties of
the "as produced" geogrids can, therefore, be used with
confidence in the predictions of wall behaviour.

For design purposes, a partial factor of safety is

proposed to take account of variations in production
quality and damage due to construction processes. For
retaining walls constructed from gravelly sand fills under
well defined conditions the manufacturer's recommended
value of this partial factor of safety appears to be very
conservative.

ACKNOWLEDGEMENTS
This work was carried out as part of the interaction
between the University of Strathclyde and Netlon Limited
within the S.E.R.C. Teaching Company Scheme.
The authors wish to thank Professor A McGown and Dr K Z
Andrawes of the University of Strathclyde for their
helpful advice during this scheme. Thanks are also due to
the staff of the Royal Military College, under Professor P
M Jarrett and Dr R Bathurst, who assisted in recovery of
the geogrids.

REFERENCES

1. Department of Transport: Specification for Road and

Bridge Works, London, 1976.

2. McGown, A, Andrawes, KZ, Yeo, KC and DuBois, D. The

load-strain-time behaviour of 'Tensar' geogrids: Polymer
grid reinforcement, Thomas Telford Limited, London, 1985.

3. Test Methods and physical properties of 'Tensar'

geogrids: Netlon Limited, 1986.
PRELIMINARY ASSESSMENT OF SIDEWALL FRICTION ON
LARGE-SCALE WALL MODELS IN THE RMC TEST FACILITY

Richard J. Bathurst and Daniel J. Benjamin

Civil Engineering Department

Royal Military College of Canada

1.0 Introduction

1.1 General
The use of laboratory models to study the behaviour of soil retaining wall systems is
common in the literature. Typically, experiments are performed in parallel-sided boxes
of finite width with one of the perpendicular faces representing the model wall under
study (e.g. Rowe, 1971). Indeed, the RMC test facility is a rather large example of this
common laboratory approach. Unfortunetly, test facilities of this type can be expected
to introduce some deviation in model response from the behaviour of the same system
with infinite width due to sidewall friction. Strategies to reduce the influence of the
apparatus edge effects include the use of friction reducing surfaces at the sidewalls and
articulated model facings. Both approaches have been employed in the tests reported
in the companion paper by Bathurst, Wawrychuk and Jarrett.
Bransby and Smith (1975) reported the results of small-scale retaining wall tests
carried out using a sand backfill and rigid facings that were allowed to rotate outwards
about the toe. The test results gave a reduction in the active earth pressure coefficient
(Ka) of 14% for a wall height to width ratio H/w = 2 and a sidewall friction angle of
4Jsw = 5.7°.
As a result of discussions held during the NATO Workshop a preliminary inves-
tigation has recently been completed to study the contribution of sidewall friction to
the performance of retaining wall models in the RMC Test Facility. The investigation
comprised the following:
a) Direct shear tests to examine the shear-deformation behaviour of the test facility
sand/polyethylene/plexiglas sidewall interface; •
b) A large-scale unreinforced wall test to determine boundary forces acting on a model
propped wall prior to and at soil failure;
c) A stability analysis of the entire block of unreinforced soil to estimate the contribu-
tion of sidewall friction to the vertical equilibrium of the soil mass after surcharging;
d) An analysis of the stability of the unreinforced test at limiting equilibrium to esti-
mate the contribution of sidewall friction to wedge stability using direct shear test
data.

2.0 Test Results

2.1 Direct Shear Tests
The results of direct shear tests carried out on samples of well-compacted dry sand
are plotted on Figure 1. The results include direct shear tests performed at RMC
(Lescoutre, 1986) using a small shear box apparatus (36cm2 in plan area) and tests
carried out at the University of Oxford using a larger direct shear apparatus (393 cm 2
in plan area). The large apparatus test results are also reported by Jewell (1987b) in
these proceedings. The combined sets of data give a peak friction angle for dense RMC
181

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 181-192.
© 1988 by Kluwer Academic Publishers.
182

80

70

60

7p...
50
~
'----'
I-
m
m
Q) 40 Shear Box Size
.....,
I-<
m t:::. 36 cm 2 (Lescoutre, 1986)
I-<
CIl 0 393cm 2 (Jewell, 1987b)
30
Q)
,...q
m 0 36cm 2
0 1m 2
20

10

60 70 80
normal stress an

Figure 1 Results of Direct Shear Tests for Dense' Sand Samples

sand of <P = 53°. Also shown on the figure are the results of direct shear tests carried
out to determine the shear-deformation response of the sidewall surface/soil interface.
The shearing plane in these tests comprised two layers of 0.135 mm thick polyethylene
sheeting over plexiglas with the sheeting left unrestrained. The data from small-scale
tests gives a peak sidewall friction angle of <Paw = 15°. This peak friction angle is lower
than the value <Paw = 20° that was determined from the result of a large (1m x 1m)
direct shear test reported in Bulletin 1 by Bathurst and Jarrett (1986) (see Figure 2).
This test was carried out under very low normal stress and was undertaken to give a
rough. estimate of the sidewall friction angle for prediction purposes and to confirm that
the polyethylene/plexiglas construction did reduce sidewall friction in the test facility.
The figure also shows that an attempt to further reduce sidewall friction by introducing
silicon oil between composite layers was not successful at very low normal stress. Based
on the comments made above it is reasonable to assume that the value <P.w = 20° is an
upper-bound estimate of sidewall friction angle for the trial walls reported by Bathurst
et al. in these proceedings.
 183

1.1

1.0

0.9

0.8

7p...
..;.::
........- 0.1 1m X 1m shearbox aQQaratus
><
~
Q)-

~
u
.....
0
'+-<
0.6 6x
.....
o:l
Q)
..d
'"
., 0.5
Cd Fx
Q
0
N
.;::
0
..d 0.4 eand - two
po lyethy lene
1interface
sheets - plexiglas
0.3
0 Test No.1 no lubricant

0.2
• Test No.2 silicone oil between
polyethylene sheets and
polyethylene and plexiglas
0.1

o
05--L~--~-L--L-----------~-------------L~
0.1 0.2 0 .3 0.4 0.5 1.0 1.5
horizontal displacement, 6x (mm)

Figure 2 Large-Scale Direct Shear Tests

184

Figure 3 shows typical load-deformation behavior of direct shear tests carried

out using dense sand samples under similar normal stress levels for sand alone and
sand/polyethylene/plexiglas tests. The figure illustrates that full sidewall friction ca-
pacity is mobilized after relatively little deformation. The data from the large test on
Figure 2 shows that only about 0.5 mm of shear displacement is required to achieve
this condition. Furthermore, both the large-scale and small-scale tests did not show
any systematic reduction in ¢sw at large deformations. The figure also shows that sand
dilation in the sand/polyethylene/plexiglas tests was not recorded. This result is con-
sistent with the visual observation that shearing in the direct shear apparatus occurred
between the layers of polyethylene sheeting. The implication to the RMC facility is that
full sidewall friction capacity will be mobilized during placement and compaction of the
RMC sand and during surcharging of the retaining wall models.

~ 40 sand/sand Pd = 1.86 (Mg/m 3 )

"
::;;;
~

f..
30
rn
rn
11)
an = 27.1 kPa
10<
+>
rn
10<
III
11)
20
~
rn
:sand/polyethylene/plexiglas Pd = 1.85 (Mg/m 3 )
10
an = 27.2 kPa

00 6 7 8
horizontal movement (mm)

4
1:J sand/sand
i=: 3
S
11)
0

~ ~ 2
S '"0

]~ O~~~~----~---L~----~----~-----L----~----~
~
11)
7 8
>
horizontal movement (mm)

Figure 3 Load-Deformation from Direct Shear Tests

 185

2.2 Large-Scale Unreinforced Test

A large-scale unreinforced test was carried out to measure retaining wall forces
and vertical earth pressures within the soil mass. Measured values have been used
to assess the influence of sidewall boundary conditions on retaining wall behaviour.
The general test arrangement is given on Figure 4. The wall was supported during
sand backfill placement as indicated on the figure. A computer-controlled actuator was
used to provide a lateral prop support to the rigid wall facing at a point 2/3 from the
base. The facing was constructed with the geogrid/wall connections in place in order
to generate facing/soil interaction in a manner similar to that expected in reinforced
systems. In addition, the wall was restrained vertically and horizontally at the toe at all
stages in the test and load cells were used to record forces at the facing panel supports.
Following construction, the soil behind the wall was surcharged to 22 kPa and the wall
forces and vertical earth pressures recorded. Next, the full width of wall facing was
allowed to rotate outwards about the toe at a controlled rate of 1.25 X 1O-4 rad/min
(i.e. 0.25 mm/min at the actuator) until an active earth pressure condition resulted
(i.e. soil failure).
The soil behind the wall was determined to have failed after a toe rotation of
about 0.01 radians or 20 mm movement at the actuator location. At failure, the total
horizontal and vertical forces acting on the wall were measured to be about IDA kN/m
and 6.7 kN/m width respectively. The magnitude of wall rotation required to generate
the active condition is similar to values reported by Smith (1972) using a dense dry
sand backfill and rigid retaining wall models.

surcharge qo

hydraulic actuator

unreinforced sand

-
backfill 3m
2m

blinding layer \.
------~
O.6m

I .. base slab
6m

- pressure cells

Figure 4 General Arrangement for Large-Scale Unreinforced Propped Wall Test

186

3.0 Results of Analysis

3.1 Vertical Stability
.The total force contributions to vertical equilibrium of the retained soil within the
test facility can be estimated from the results of the earth pressure cells located at the
bottom of the test facility and vertical force measured at the bottom of the propped wall.
The distribution of vertical earth pressures recorded below a portion of the retained soil
mass after the soil was surcharged to 22 kPa can be related to Figure 5a. The figure
shows the fraction (R) of surcharge pressure and soil self-weight recorded by the earth
pressure cells where:
(1)

and qb denotes base pressure. It is clear from the figure that losses occur. These
losses are pronounced at the wall facing due to the relatively high wall/soil interaction
developed at this interface (i.e. 6 = 33°) and are diminished further from the facing
panels. A reasonable average R factor is about 0.88 or 0.9 indicating that about 10
to 12% of vertical earth pressure at locations well within the block of soil is lost to
sidewall friction. It is possible that a portion of this under-registration is also due to
cell/soil interaction. It should be noted that the surcharge pressure (qo = 22kPa) in
these calculations has been corrected for the influence of airbag coverage which does not
extend completely to the sides of test facility boundaries when the airbags are inflated.
The reduction in total soil self-weight that is recorded at the base of the test facility
can be denoted by X sw . In addition, there will be a portion of the total surcharge force
that can be expected to be carried by the sidewalls due to sidewall friction. This
component of total vertical force is given the term X . Consider first the contribution
of sidewall friction generated due to soil self-weight: the unit sidewall friction fsw can
be expressed as:
faw = Kswqg tan <Psw (2)
Here qz = "(Z and is the vertical ·stress due to soil self-weight acting at depth z below
the top of the wall and Ksw is the coefficient of sidewall earth pressure. Integration of
equation (2) over both sidewalls gives a total sidewall force due to soil block self-weight
according to:
Xew = Ksw"(H2Ltan<psw (3)
where L is the length of the soil block and H is the soil block height. The additional
sidewall friction generated due to surcharge loading can be accounted for by adopting an
approach similar to that proposed by Jewell (1987b) in this proceedings. The attenuated
vertical stress q" acting at depth z can be described by the relationship:

(4)
where:
C 1_- 2Ksw tan <Psw (5)
W

and w is the width of the soil block. The total surcharge force carried by the test facility
sidewalls becomes:
(6)
Assum:ing that the vertical force acting on the lagging board wall at the back of the test
facility is equivalent to that measured on the front wall, it is possible to calculate the
value of Ksw required to satisfy vertical equilibrium. The results of these calculations
are shown on Figure 5b for <Pew = 15° and 20°. The figure shows that Ksw varies from
about 0.23 to .52 for the range of Rand <Psw considered. This range includes the value
of 0.4 that Jewell (1987b) estimates to be appropriate for the RMC sand under near
plane-strain conditions.
 187

distance from front face (m)

a) Distribution of Vertical Earth Pressures at Soil Base

0.6 r - - - -..........-----,.....----.----~

0.5

~
rn
~ 0.4

0.3

0.2 "--_ _.........._ _ _..L-_ _ _ _ _ _ _ _.....

0.88 0.89 0.90 0.91 0.92
R
b) Coefficient Ksw versus Rand <Psw

Figure 5 Results of Vertical Stability Analysis

188

3.2 Wedge Stability at Limiting Equilibrium

A simple Coulomb wedge approach can be used to make an assessment of the
contribution of sidewall friction to model wall stability at limiting equilibrium.
The horizontal and vertical equilibrium of a critical soil wedge subject to sidewall
tractions can be related to the system of forces shown on the diagram in Figure 6. In
this system the total sidewall resisting force due to wedge self-weight is again denoted
as Xsw but is assumed to act parallel to the plane of soil failure. As in the previous
analysis, sidewall friction can be expected to reduce the effect of any surcharge qo acting
at the surface of soil. This mechanism can be accounted for by introducing a vertical
force vector X acting in the opposite direction to the wedge self-weight vector W. If
the influence of sidewall friction is neglected the solution to this class of problem can be
found in soil mechanics textbooks. The orientation of the critical failure plane in these
problems is at (or close to) 0 = (1r/4 + r/>/2).
Consider first the contribution of soil self-weight to sidewall friction: Integration
of equation (2) over both sidewalls gives a total sidewall force due to wedge self-weight
according to:
(7)
The additional sidewall friction generated due to surcharge loading can be accounted
for by adopting the approach proposed by Jewell (1987b) in this proceedings. The
attenuated vertical stress qz acting at depth z can be described by the relationship:

(8)
where:
2Ksw . (1r
C 2 = --tanr/>swsm - - 0) (9)
w 2
Equations (2), (7) and (8) lead to an expression for the total surcharge force taken by
the sidewalls:
Xq = C 2 wqo tan(~ - 0)
2 i0
H (H - z)e- C2Z dz

The total active force P A acting on the wall can now be expressed as:
(10)

W + qoBw - Xq - Xsw[sin(O) + ~ cos(O)]

PA = -----~~-~-~-~---~ (11)
sin(6) + ~ cos(6)
Here:
A = cos(O) + sin(O) tan(r/»
1r
B = Htan("2 - 6)
(12)
D = sin(O) - cos(O) tan(r/»
W =-yHBw/2
The cri.tical solution can be determined numerically by varying 0 to find dP A/ dO = o.
With the exception of the coefficient of sidewall earth pressure Ksw, all variables
in equation (11) are known or can be estimated with some confidence from tests carried
out at RMC. Various solutions to equation (11) in terms of horizontal and vertical com-
ponents of P A are given in Figure 7. Values of P Ah = P A cos(6) and P Av = P A sin(6)
are plotted against assumed values of Ksw using a range of soil friction angles r/> and
 189

sidewall friction angles <Psw = 15 and 20°. The wall facing/soil friction angle 8 was cal-
culated directly from the measured wall forces at failure. Superimposed on the figure
are the measured ranges of wall forces acting at limiting equilibrium in the large-scale
unreinforced wall test. The Ksw = 0 condition represents the ideal situation of no side-
wall friction. As may be expected, the measured facing loads fall below these values
indicating that sidewall friction does occur. The lateral confinement of the sand soil in
the RMC test facility suggests that the fully-mobilized soil friction angle is somewhat
higher than the value of 53° measured in direct shear tests (e.g. Rowe, 1969). Jewell
(1987a,b) suggests that a value of 55 or 56° may be appropriate. If this is the case,
the percentage reduction in P A due to sidewall friction is of the order of 12 to 18%.
For <p = 55° and sidewall friction angles of <Psw = 15 to 20° the corresponding range for
Ksw is 0.22 to 0.5. This range of values is in agreement with the results of the vertical
stability analysis described earlier and the value of 0.4 proposed by Jewell (1987b).

rz

()

Figure 6 Coulomb Wedge Analysis with Contribution of Sidewall Friction Forces

190

no sidewall friction condition

40
cPsw = 20°
cPsw = 15°
1= 18kN/m 3
{) = 33°
35 qo = 22kPa
Z
6 ......
..e cP
«
0... 30 }50

25 }53

}55

20 0 0.3 0.5
Ksw

30
cPsw = 20°
cPsw = 15°
1= 18kN/m 3
25 {) = 33°
qo = 22kPa
~

Z cP
..!<:
20
~

>
« }50
0...

15
measured range of wall force at failure

10 0
0.1 0.2 0.3 0.4 0.5
Ksw

Figure 1 Results of Stability Analysis and Measured Wall Forces

 191

4.0 Summary of Results and Implications to RMC Trial Walls

The results of the laboratory tests and analyses carried out to investigate sidewall
friction contribution to test walls constructed in the current RMC test facility suggest
that sidewall friction is fully-mobilized at all stages in construction and surcharging. The
fully-mobilized friction angle is ¢sw = 15° and can be assumed to operate at large defor-
mations including soil failure. The results of stability calculations assuming ¢sw = 15 0
show that a reasonable value for the coefficient of sidewall earth pressure Ksw is 0.4
which is in agreement with the value proposed by Jewell (1987b) based on the results
of laboratory plane-strain tests. Measurements taken during surcharging of the unrein-
forced wall to qo = 22 kPa showed that the sidewall friction effect is responsible for a
reduction of about 10 to 12% in vertical earth pressure at the base of the test facility.
The limiting equilibrium wedge analysis was shown to give a correct prediction
of wall forces for reasonable values of system parameters. Stability analyses can be
performed for a hypothetical unrein/orced test carried out in the RMC Test Facility
with and without sidewall friction and subject to a surcharge pressure of qo = 50 kPa.
Assuming the following parameter values: H = 3m, 'Y = 18 kN 1m3 , 8 = 33°, Ksw = 0.4
and ¢ = 55° the analyses show that the active earth pressure force PAis reduced
below the ideal no-friction condition by 14% for ¢sw = 15°. If it is assumed that the
sidewall force contributions calculated for this unreinforced case are applicable to the
same soil mass in a reinforced condition then, the 14% reduction in P A can be used as a
preliminary estimate to account for the additional capacity of the RMC trial walls due
to sidewall friction.

Acknowledgements
The authors would like to acknowledge the contribution of J. Bell (Research As-
sistant) who helped with many aspects of the test program reported here. The authors
would also like to thank J. DiPietrantonio who drafted the figures and Dr. P.M. Jar-
rett for his input during this investigation. Funding for this investigation was provided
through the ARP program and the Chief of Construction and Properties, Dept. of
National Defence (Canada).
192

References

1. BATHURST, R.J. and JARRETT, P.M. (1986)

Class A Prediction Exercise for Reinforced Earth Walls,
Bulletin No.1 for NATO Advanced Research Workshop, Application
of Polymeric Reinforcement in Soil Retaining Structures
Departments of Civil Engineering RMC and the University of Strathclyde
2. BATHURST, R.J., WAWRYCHUK, W.F., and JARRETT, P.M. (1987)
Laboratory Investigation of Two Large-Scale Geogrid Reinforced
Soil Walls .
Application of Polymeric Reinforcement in Soil Retaining Structures
NATO Advanced Research Workshop, Royal Military College of Canada
June 1987
3. BRANSBY, P.L. and SMITH, LA.A. (1975)
Side Friction in Model Retaining-Wall Experiments
Journal of the Geotechnical Engineering Division, ASCE,GT7
4. JEWELL, R.A. (1987a)
Reinforced Soil Wall Analysis and Behaviour
A.pplication of Polymeric Reinforcement in Soil Retaining Structures
NATO Advanced Research Workshop, Royal Military College of Canada
June 1987
5. JEWELL, R.A. (1987b)
Analysis and Predicted Behaviour for the Royal Military College Trial Wall
. Application of Polymeric Reinforcement in Soil Retaining Structures
NATO Advanced Research Workshop, Royal Military College of Canada
June 1987
6. LESCOUTRE, S.R. (1986)
The Development of a Large-Scale Test Facility for Reinforced
Soil Retaining Walls
M.Eng. thesis, Royal Military College of Canada, Kingston
7. ROWE, P.W. (1969)
The Relation Between the Shear Strength of Sands in
Triaxial Compression, Plane Strain and Direct Shear
Giotechnique 19, No.1, pp 75-86.
8. ROWE, P.W. (1971)
Large Scate Laboratory Model Retaining Wall Apparatus
Proceedings of the Roscoe Memorial Symposium, Cambridge University,
March 1971
9. SMITH, LM. (1972)
Discussion, 5'th European Conference on Soil Mechanics and
and Foundation Engineering, Madrid, 1972
 ANALYSIS AND PREDICTED BEHAVIOUR
FOR THE ROYAL MILITARY COLLEGE
TRIAL WALL
R.A. JEWELL
University of Oxford

Abstract
Predictions are made for the incremental and propped reinforced
soil walls built at the Royal Military College, Kingston, and
these illustrate the theory and analysis presented in a
companion paper. New ideas are applied to the interpretation
of shear box data to derive parameters for the sand, and
supplementary test data are reported. The predictions are
based on two "bounding" equilibrium states for reinforced soil
walls, and the influence on these of the parameters and
boundary conditions in the trial walls are investigated. The
side wall friction appears to have a major influence: a closed
form analysis is presented for side wall friction so that it
can be taken into account for both the self weight and sur-
charge loading. The predictions are then presented and
compared with the measured data. As all aspects of the wall
behaviour under different loading conditions were predicted
this provides a comprehensive test of the analysis. The find-
ing is that the analysis captures almost all the main features
of the observed behaviour, particularly the patterns of behav-
iour, but also the magnitudes. Further, it provides a natural
link between the results of the two different trial walls.

INTRODUCTION
1.1 Introduction
This is a companion paper to "Reinforced soil wall analysis and
behaviour" by R. A. Jewell published in the same proceedings,
and hereafter referred to as the companion paper. Although all
the theoretical concepts and the analysis are described in the
companion paper additional details about the theory are brought
out by the prediction for the RMC trial wall.
The data for the trial wall were presented in Bulletin 1 by
Bathurst and Jarrett (1986).

193

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 193-235.
© 1988 by Kluwer Academic Publishers.
194

1.2 Organisation of the predictions

The predictions for the incremental trial wall, built at the
Royal Military College (RMC) in July 1987, are set out in the
following manner.
First the properties of the materials used for the construction
are examined and parameters evaluated for the subsequent
analysis. Additional test data for the sand are presented from
tests carried out independently at Oxford University.
The geometry of the trial wall and the theoretical distribution
of the soil stresses due to self weight and surcharge loading
are then discussed. The allocation of the required soil
stresses to the individual layers of reinforcement is
described.
A complete set of predictions for the central loading case
(surcharge 12 kPa, duration 100 hours) is then made assuming
ideal boundary conditions - that means no benefit is allowed
from the boundaries of the wall, importantly the rough rigid
base, the interaction between the facing and the base and the
side wall friction in the test wall.
The additional outward movement induced by the construction of
the incremental wall is then calculated for purely translating
and purely rotating facing panels. This allows the overall
face movements for the wall to be estimated, to complete the
set of predictions.
The likely influence of the actual boundary conditions in the
trial wall are then discussed. The rough, strong base and the
connection between the face and the base are both likely to
reduce the reinforcement force magnitude and distribution in
the lower reinforcement layers. The choice is made to ignore
these two beneficial factors because of the lack of an analysis
or basic data for their evaluation.
The influence of the measured side wall friction is anticipated
to be significant, particularly with surcharge loading. A
simple closed form aI1alysis is derived to allow side wall
friction to be taken into account.
The predictions for the wall are then repeated allowing for
side wall friction. For this analysis the (predicted) most
likely shearing resistance for the sand is used, and the
calculations carried out for the central loading case and for
the end of the test (surcharge 50 kPa, duration 1000 hours).
Finally the strain compatibility in the trial wall is examined
using the stress-strain properties for the sand deduced from
the direct shear tests. The evaluation of this important
aspect of reinforced soil behaviour is dealt with at the end of
the paper because there are no direct data for the
stress-strain characteristics of the sand. The interpretation
 195

for direct shear test data proposed in the companion paper is

illustrated and used with an elastic plastic model for the
sand.
1.3 Measured behaviour
In preparing the final manuscript for publication, after the
trial wall has been built, it has been possible to include a
section at the end of the paper commenting on the measured
behaviour compared with the predicted behaviour. A general
report on the predictions is published elsewhere in the
proceedings.

SOIL MATERIAL PROPERTIES

2.1 Soil shearing resistance

The two important values of frictional shearing resistance are
the peak angle of friction and the critical state angle of
friction. The mobilised angle of friction in the reinforced
soil wall will probably be in this range. The magnitude of
tensile strain in sand bslow the critical state stress ratio on
the stress paths relevant for a reinforced soil wall is small
relative to the working tensile strain in polymeric
reinforcement.
The peak shearing resistance depends on the combination of mean
stress (pressure) and specific volume (density) in the soil.
The pressure and density combinations for the sand fill in the
trial wall will be estimated to allow the appropriate peak
shearing resistance to be evaluated.
The shearing resistance also depends on the loading conditions
and stress path in the soil. For the incremental wall there is
a loading stress path (increasing mean stress and stress ratio)
under plane strain conditions. The direct shear test data
which are available for the sand are relevant to these
conditions.
For the propped wall there is an unloading stress path
(decreasing mean stress and increasing stress ratio) under
plane strain conditions. The standard direct shear test does
not follow such a stress path. A biaxial test with a constant
vertical applied stress and reducing lateral stress would
provide more relevant data.
2.2 Material mineralogy, grading and density limits
The data are provided in bulletin 1 except where stated
(Ba thurst and Jarrett, 1986). The sand is a mix of feldspar
and quartz, with relatively angular particles between medium
sand and fine gravel size. The measured maximum and minimum
dry densities were 19.2 kN/m3 and 15.9 kN/m3.
196

The (dry) density for the compacted sand in the trial wall is
expected to be 17.6 kN/m3, a relative density of 52%. The bulk
density of the slightly moist sand with the anticipated 3%
moisture content has been taken as 18 kN/m3 for the prediction.
Because the soil is relatively coarse grained the low moisture
content will probably not significantly influence the effective
stresses in the soil. The properties for the dry sand were
used for the prediction.

2.3 Critical state ,angle of friction

One measurement for the plane strain critical state angle of
friction comes from direct shear tests at large shear
displacement, although the measurement is imprecise at large
displacement. Five of the direct shear tests repqrted in
bulletin 1 were sheared to sufficiently large displacement.
The average critical state angle of friction from these tests
was ¢" = 38°. The results are included in Table 1.
Four additional direct shear tests on the RMC sand were carried
out independently at Oxford University, as described in the
next section. Three tests in a larger shear apparatus gave an
average measured critical state angle of friction ¢" = 38,5', and
the one test in a standard direct shear apparatus gave ¢" = 38°.
The link between soil mineralogy and the critical state angle
of friction has been highlighted recently in Bolton's (1986)
correlation of data for sand. Bolton indicates that the
presence of feldspar in an otherwise quartz sand gives ¢,," 36°
to 37°. For pure felspathic sand the critical state angle of
friction is higher still ¢,," 39' to 40°, Koerner (1968). Hence
a critical state angle of friction between 36° and 40° is
consistent with the mineralogy of the RMC sand.
Cornforth (1973) suggested a simple approximate measurement for
the critical state angle of friction. He proposed that the
angle of repose of a loosely tipped slope of dry sand subject
to excavation at the toe would approximately equal ¢". This
test was carried out at 'Oxford with' the dry, RMC sand, in a
plane strain glass tank 150 rom wide, and gave an angle of
repose 38' ± 1
0.

Critical state strength for the RMC sand

The conclusion is that the feldspar content of the RMC sand
results in a higher critical state angle of friction than the
typical values for quartz sand. The plane strain critical
state angle of friction for the RMC sand is likely to be in the
range ¢" = 37' to 39° , and the mean value was used in the
predictions.
 197

Table 1. Analysis of the direct shear test results given

in bulletin l-

Ory density :kN/m3 16.6 17.5 18.3 18.3

Relative density :% 21 52 73 73
Vertical stress :kPa (¢ ds) p (¢dS) p (¢ dS) p (¢ ds Lv

l3 .5 46.5 53 56.5 38.5

27.1 42 50 54 35.5
40.6 46 47 53.5 41
54.2 43.5 49 54 38
67.7 43 45 50 38.5

2.4 Peak angle of friction

Standard (small) direct shear tests

Fifteen standard (small) direct shear tests on the RMC sand
were reported in bulletin 1. These comprise five tests under
different stresses, each repeated at three relative densities.
The data have been analysed to give the peak (secant) angle of
friction for the sand as a function of the relative density and
the applied vertical stress, Table 1.
60
~ RIoIC Small, RD 7~

Z .I- RIoIC Smal~ RD 52lC

0 -~
~ ~- •
§ 58
~~-~
RIoIC Small, RD 21"
J..arve.
if l( OXF RD 55lC
~...,...

-- --
~ ~~- ~ ~
v
"1-- OXF Small, RO 150"
La..
0 52 -)to-.
~~~~~
x

-
_~

+_x ~-

-----
~ ~ ~.

C> + "
~ 48
+
~J:
~.~-----
--- • +
Ul
44 ---------------~ --.J!. __

~ •
0:: <40
i5 &tImated critical IItcIte
~'
a.' 36
i 12 kPa can t50 kPa can
2 2.4 3.2 3.11 4 4.4

VERTICAL STRESS (Natural Logarithm) :kPa

Fig. 1 Variation of the peak direct shear friction angle with
pressure and density, for the RMC sand.
198

Frictional shearing resistance above the critical state angle

of friction is derived from the sand dilatancy, which is
related to the density of the soil and the mean stress level.
To illustrate the relationship, the peak, direct shear angles
of friction in Table 1 are shown plotted in Fig. 1 as a
function of the relative density of the sand and the logarithm
of the applied vertical stress.
The peak, direct shear angle of friction can be selected from
Fig. 1 for any des.ired combination of relative density and mean
pressure in the soil within the range of the test data.
Direct shear tests at Oxford
The ratio of the length of the test specimen to the mean
particle size is about 45 for the RMC sand in a standard direct
shear apparatus. To check for possible scale effects, three
direct shear tests were carried out at Oxford University in a
254 mm long, 152 mm wide and 152 mm deep direct shear
apparatus. (Jewell and Wroth (1987) give details of the
apparatus} .
The conventional test arrangement was adopted (to be comparable
with the RMC tests) with a rough, rigid top platen not secured
to the top half of the apparatus. The tests were on dry sand
with samples prepared by tamping in five layers with a 1.5 Kg
tamper.
The test results are given on Fig. 2, and the measured peak and
critical state angles of friction, and the peak angles of
dilation are reported in Table 2.
One standard (60 mm by 60 mm) direct shear test was also
carried out at Oxford University for comparison purposes with
the three larger tests, and the fifteen RMC tests. The result
is also shown in Fig. 2, and the measured parameters included
in Table 2.
The measured peak, direct shear angles of friction in the
Oxford tests are shown "in Fig. 1 and compare well with the
previously reported results. This provides confidence in the
direct shear strength data for the RMC sand.

Table 2. Results from the direct shear tests at Oxford.

Relative density a, (¢ •• ) p (¢ .. )" Cdyldx)max Apparatus
:% :kPa :mmxmm
55 17.5 52 38 .33 2S4xlS2
55 32.5 51 39.5 .35 2S4xlS2
55 47 51.5 38 .30 254xl52

50 44 52.5 38 .40 60x60

 199

(I)

13 1.4
~ 1.2
~
~ 0.8

"13
(I) 0.6
• OXF La/oge, 17.5 lcPa
0.4 . OXF l.arge, 32.5 lcPa
~ 0 OXF l.arge, 47 kPa
0.2
~
::t:
I> OXF Small, 47 KPa
(I) 0
0 2 4 6 8 10 12
(0) SHEAR DISPLACEMENT x:mm

E 2.2
E
>.
1.8
~
a 1.04-
~
a.
~
~ 0.6

~ O~~~~~________________________~
2 6 6 10 12
(b)
SHEAR DISPLACEMENT x:mm
Fig. 2. Direct shear tests on the RMC sand completed at Oxford
University.

Peak, direct shear angle of friction

The expected relative density for the compacted sand in the
trial wall is 52% (section 2.2). The vertical stress at half
the depth in the wall with the 12 kPa and 50 kPa surcharge
loads will be approximately (0,),,= 40 kPa and (o,)so= 80 kPa.
Assuming a (conservative) value of the active earth pressure
K. = 0.25 , the estimated mean stress in the sand for the two
loading cases is S~= 25 kPa and Sm= 50 kPa.
The peak, direct shear angle of friction values for these
combinations of relative density and mean stress level are,
from Fig.!., (¢d')"= 51° and (¢d.)SO= 49'.
Plane.strain angle of friction
The plane strain angle of friction, which is the relevant value
for the prediction, can never be smaller than the direct shear
angledf friction deduced from the standard interpretation for
a standard direct shear test (see the companion paper). How
much larger is it likely to be?
200

One approach is to select the boundary conditions for a direct

shear test to constrain the sand to deform in as uniform a
manner as possible across the centre of the apparatus. For
dense sand this may be achieved by fixing the top loading
platen to the top half of the shear box, forcing the top half
of the apparatus to move as a unit, symmetrically with the
bottom half. The relative displacement measurements in the
test should then more closely represent the angle of dilation
in the sand. This is discussed by Jewell and Wroth (1987).
The measured angle of dilation and the critical state angle of
friction for the sand can then be combined using a flow rule
to estimate the plane strain angle of friction. It will be
proposed in a future publication that this test arrangement
and accompanying analysis will usually be conservative, only
at best giving the plane strain angle of friction.
One test with these boundary conditions was performed at
Oxford University on the RMC sand in the 250 rom by 152 rom
shear apparatus. The sample density was 18.2 kN/m3, a
relative density 70%, and it was tested with a vertical stress
30 kPa. The test results are given in Fig. 3. The measured
variation of the direct shear angle of friction throughout the
test is shown together with the plane strain angle of friction
calculated from Bolton's (1986) simple flow rule
¢ ps = ¢ cu + 0 , 81/.1 ( 1)

with the measured angle of dilation and the expected critical

state angle of friction for the RMC sand (section 2.3).
These results indicate a peak, plane strain angle of friction
for the RMC sand of the order q'Jps" 55' to 56'. In other words,
the plane strain angle of friction is likely to be at least 5'
higher than the measured direct shear angle of friction.
Selected friction angles
The discussion in the companion paper suggests that the
mobilised stress ratio in the sand fill in the trial wall is
likely to be high and close to peak.
A range of mobilised friction angles are used for the first
set of predictions with ideal boundary conditions to indicate
the significance of the angle of friction. The range used is
q'J=45', 50' & 55°.

Lower mobilised angles of friction are adopted for the

calculation of the construction induced movements, because
these occur with lower strain in the soil. The values
q'J = 30°, 35° & 40° were selected.

For the actual prediction of the wall behaviour allowing for

the expected boundary conditions a mobilised plane strain
angle of friction q'J= 50° was selected. This is below the peak,
plane strain angle of friction for the sand.
 201

60

III
50
0'1
Q)
"U
40
z
0
~
0:: 30
lL.
• Direct shear result
lL.
0 A Angle of dilation
w 20 Flow rule, 39 c.l/.
...J
0 Flow rule, 37 c.y.
Z
«
10

0
0 2 4 6 8 10 12 14
SHEAR DISPLACEMENT x:mm
Fig. 3. Direct shear test on the RMC sand with enforced
symmetrical deformation. Interpretation for the plane strain
angle of friction using a flow rule.

The same friction angle was adopted 'for both the 12 kPa and
the 50 kPa surcharge load cases. This was because the
addi tional soil strain under the higher surcharge is
anticipated to cause additional soil shearing resistance to be
mobilised in the sand counterbalancing (to well within the
accuracy of the prediction) the influence of the increased
mean stress level in the soil.
The angle of dilation corresponding with the mobilised angle
of friction is ~~ IS·, from Bolton's flow rule and the
estimated critical state angle of friction.
2.5 Estimated soil shear modulus for compatibility
A simple elastic plastic model for the sand is proposed for
making estimates of strain compatibility. An interpretation
for the direct shear test has been presented to quantify the
relationship between the mobilised soil frictional resistance
and the principal tensile strain. The relevant equations were
derived in the companion paper.
The shear modulus for use with the compatibility curve is
estimated from the stage of the direct shear test between the
critical state stress ratio, marking the end of the
substantial rotation of principal axes at the beginning of the
test, and the selected mobilised plane strain angle of
friction.
202

The shear displacement developed between the critical state

stress ratio and the anticipated plane strain mobilised stress
ratio (~=50') is approximately 1mm, Fig. 3. The sample is 150
mm thick.
The change in stress ratio is
L1(tls) = sin t/> - sint/>cv = sin50° - sin 38 0 = 0.15 (2)

The estimated shear strain is

OX 1
L1(y)=-=-=7 10- 3 (3)
H 150

The estimated shea~ modulus for a mobilised angle of friction

50'is
L1(t/s)
( Gis) = --= 23 (4)
50 L1(y)

The initial shear modulus is approximately twice as large,

Fig. 3, giving a range for use with the compatibility curve
(Gis) = 20 to 50.

The magnitude of principal strain required to reach the

mobilised angle of friction ¢ = 50', with a shear modulus
(Gis) = 23, is

() L1(t/s) 3 10- 3
E 3 50 = 2 ( Gis) (5)

The analysis indicates about 0.3% tensile strain required to

mobilise the working shearing resistance in the sand.
Selected shear modulus
The selected shear modulus for the sand to allow for an
analysis of strain ccmpatibility is in the range (G/s)= 20to 50.
This indicates an anticipated principal .tensile strain
required in the sand to mobilise an angle of friction ¢= 50' of
the order f3~ 0.2% to 0.4%.
 203

REINFORCEMENT MATERIAL PROPERTIES

3.1 Reinforcement stiffness

Preliminary estimates suggest that the wall will achieve
equilibrium with a reinforcement force in each layer of the
order 5 kN/m or less. No assessment is made, therefore, for
the reinforcement strength characteristics as loss of
stability by reinforcement rupture is not in question.
The isochronous load extension curves at 20'C for the
"ex-works" Tensar SR2 reinforcement given in bulletin 1 (Fig.
7c) are the correct data from which to estimate the
reinforcement stiffness for the appropriate durations of
sustained load. The expected design temperature 20' to 22'C' and
mechanical damage to the reinforcement material during
construction may reduce the stiffness somewhat, but this
should be negligible in the trial.
The isochronous load extension data has few measurements at
lower loads, having been derived with higher reinforcement
forces in mind than those in the RMC trial walls. The lowest
test load reported for the Tensar SR2 is 13.2 kN/m, and the
lowest measured strain is at this load after 1 hour, and is
just less than 2%. The isochronous stiffness data which are
also given are for 10% extension.
The 100 hour isochronous curve was used to derive the
reinforcement stiffness for the central loading case (12 kPa
for 100 hours) even though reinforcement lower in the wall
would have been loaded for longer. Similarly the 1000 hour
curve was used for the final loading case (50 kPa after 1000
hours) as this was the longest time for which load extension
data is given. No attempt was made to extrapolate to the
stiffness at the expected cumulative time 2300 hours for all
the loading steps through the trial to the end of the test.
Stiffness values were determined from the extensions at a load
5 kN/m after 100 hours and 1000 hours. The magnitude of
extension was 0.8% and 0.9% respectively (Bulletin 1, Fig.
7c), which gives reinforcement stiffness values KlOo= 625kNIm
and K 1000 = 550 kN 1m.

Selected reinforcement stiffness

The selected values of reinforcement stiffness for the two
loading cases are K 100 = 625 kN 1m and K 1000 = 550 kN 1m.
The K 100 stiffness was also used for estimating the movements
induced by incremental' construction, although a shorter
loading ,period would have been more appropriate. The
approximate nature of the calculation for incremental
movement, however, does not justify the greater refinement.
204

GEOMETRY OF THE TRIAL WALL AND LOAD DISTRIBUTION TO THE

REINFORCEMENT

4.1 Local equilibrium

The concept of reinforcement layers maintaining local
equilibrium in the reinforced soil has well as overall
equilibrium was discussed in the companion paper. The central
notion is that each reinforcement layer provides the required
horizontal stresses in the soil for half the spacing to the
next · reinforcement layer above and below.
The geometry for the incremental trial wall is shown in Fig.
4a. For convenience the surcharge is assumed to be applied
exactly on the top of the wall at 3.0 m height. There are
four 0.75 m facing panels and the reinforcement is attached at
the lower third point· of each, 0.25 m above the panel base.
The vertical depth of soil locally supported by each
reinforcement layer in metres is 0.625, 0.75, 0.75 and 0.875
for layers 1 to 4 respectively, Fig. 4a.

0·875 m

0·75 m
3m - - - 3 m
Layer 2
0·75m

__JL~La~y:er:I::::::::~~~0.625m

(a) Reinforcement spacing (b) Self weight loading (c) Surcharge loading

Fig. 4. Local equilibrium for the reinforcement layers in the

RMC trial wall.

It is interesting to look at the percentage of the gross

required force that each layer must support with ideal
boundary conditions. The horizontal required soil stress
distribution is triangular for the self weight loading, and
uniform for the surcharge loading, Figs. 4b and c. The
percentage of the load supported by each reinforcement layer
is given in Table 3. This shows that the self weight loading
 205

is unevenly distributed to the lower layers (as expected with

uniform spacing) but that the surcharge loading is quite
evenly distributed to the four reinforcement layers.

Influence of the boundaries

The pattern of loading will be shown later to change quite
significantly when side-wall friction is included.
Meanwhile, there is a estimate for the influence of the base
boundary in the trial. If the rough, strong base boundary to
the wall acts like a strong reinforcement layer it would
reduce the load taken in the lowest reinforcement layer 1. In
the incremental wall the base would hold the 0.125 m of soil
above it in equilibrium. The required soil stresses in this
zone amount to 8% of the total for self weight loading and 4%
of the total for surcharge loading. The resulting percentage
reduction in the force in the lowest reinforcement layer 1
would be 22% and 19% for the self weight and surcharge loading
cases respectively. This is a significant change.

Table 3. Percentage of the gross required force

carried by each reinforcement layer: ideal
boundary conditions.

Loading by: Self weight Surcharge

Layer 4 9 % 29 %
Layer 3 21 % 25 %
Layer 2 33 % 25 %
Layer 1 37 % 21 %
Total 100 % 100 %

PREDICTION WITH IDEAL BOUNDARY CONDITIONS

Central loading case: 12 kPa surcharge applied for 100 hours

5.1 Introduction
The aim is to predict the reinforcement forces and force
distribution, estimate the corresponding reinforcement strains
and hence the overall elongation in each reinforcement layer
at the end of the loading period. This will be completed for
three mobilised soil strengths covering the likely range for
the trial wall. This will indicate the sensitivity of the
predicted behaviour to the soil strength properties.
206

The additional outward movements caused by the incremental

construction are estimated in section 6.4, and these allow the
overall outward movement of the incremental wall face to be
derived.
Two idealised equilibrium states will be analysed, the ideal
length and truncated length cases described in the companion
paper.
5.2 Force in the reinforcement
The maximum force in each reinforcement layer is uniform in
the zone between the most critical mechanism and the wall face
and is the same magnitude for both equilibrium cases. The
required horizontal soil stress for equilibrium in this zone
is given by the standard Rankine active earth pressure
coefficient for a smooth wall.

Table 4. Maximum force in the reinforcement in kN/m:

12 kPa surcharge, ¢= so· , ideal boundary conditions.

Loading by: Self weight Surcharge Total

Layer 4 0.95 1.36 2.31
Layer 3 2.21 1.17 3.38
Layer 2 3.47 1.17 4.64
Layer 1 3.90 0.98 4.88

Total 10.53 4.68 15.21

The calculated forces in each reinforcement layer developed by

the self weight loading and the surcharge loading for ¢= so· (
K.= 0.13) are reported in Table 4. The distribution of the
gross required force between the four reinforcement layers in
the wall is illustrated in Fig. 5.
The earth pressure coefficients for the other two soil
strengths in the range are K. = 0.17 and 0.10 for if! = 45' and 55°
respectively. The magnitude of the calculated reinforcement
forces are also illustrated in Fig. 5.
The range for the anticipated mobilised soil strength changes
the predicted reinforcement forces by about 50%.

5.3 Strain in the reinforcement

The reinforcement stiffness for the central loading case is
K 100 = 625 kN 1m (section 3.1). The maximum reinforcement strain
is linearly related to the maximum reinforcement force by the
reinforcement stiffness, so that the set of predicted maximum
strains can be shown with the addition of a new scale at the
top of Fig. 5.
 207

Maximum Reinforcement Extension

.5% 1% 1.5%
3~------------~-------------+-------------4--.

E
2.6
• 45 degrees
50 degrees
0:::
W 55 degrees
~ 2.2
C>
z
is 1.8
z
:::i
CD
w 1.4
6
CD
~
l-
I
C>
iii 0.6
I

o 2 4 6 8 10

MAXIMUM REINFORCEMENT FORCE Pr :kN/m

Fig. 5. Maximum force in the reinforcement layers: 12 kPa

surcharge, ideal boundary conditions.

The range of predicted maximum reinforcement strain is 0.3% to

1.0%, well below the 10% limit strain for the' Tensar SR2
reinforcement.
5.4 Stressed reinforcement length
Two idealised states of equilibrium in a reinforced soil wall
are examined. Neither is likely to represent the actual case
but they are likely to bound it. The ideas are described in
the companion paper where the equations and design charts are
given.
The two equilibrium states considered are the ideal length and
the truncated length cases. The zone in the soil where
reinforcement force is required for these two cases, and the
position of the most critical mechanism through the toe, have
been calculated for the three soil strengths and are shown in
Fig. 6.
208

(a) ( b) (c)

(d ) (e) (t)

Fig. 6. Predicted zones of reinforcement force for (a to c)

the ideal length and (d to f) the truncated length cases.

One point from Fig. 6 is that the location of the required

reinforcement force all lies well within the available
reinforcement length in the trial wall. This fact, combined
with the relatively small reinforcement force magnitudes in
the wall, suggests that bond between the reinforcement and the
soil is not a problem, and relative slippage between the soil
and the reinforcement in the trial wall is unlikely.
There is another implication from the above, which is that
there is unlikely to be movement in the wall due to
deformation in the soil behind the reinforced zone. As there
is a rigid foundation beneath the trial wall, this implies
that the movement at the face of the wall will be caused only
by the elongation in the reinforcement layers and the
additional outward movement resulting from the incremental
construction.
5.5 Reinforcement force distribution
The equations giving the distribution of force along the
reinforcement in the ideal length case are given in the
companion paper. The force is uniform towards the face and
decreases gradually between the most critical mechanism and
the locus of zero required force. The magnitude of the maximum
force in each layer was calculated in section 5.2.
 z 0.6r-------------, z 0.6,------------,
~
15 LAYER 1 ~ 0.5 LAYER 3
0.5 15
~ ~
C/l
0.4 ~ 0.4
~ 0.3 V ~ 0.3.r-------V'l
t'0::3 /
~
0.2 1/
V 0::0.2
~
~
~ 0.1 V ~ 0.1
~ O±-lk''":+:-..,...,,:::-r-~r_:_r:~,.........,....,.....,...,,-4 0:: 0 f--.-:T7..,...",-=-"-~~cr-:,.__._:::r:_r_::""::_l
o 0.4 0.8 1.2 1.6 2 2.4 2.8 o 0.4 0.8 1.2 1.6 2 2.4 2.8
DISTANCE FROM THE WAlL FACE :m DISTANCE FROM THE WAlL FACE :m
z 0.6,-----------, z 0.6r-------------,

~ 0.5
LAYER 2 ~ 0.5
LAYER 4
~ 0.4 ... 0.4
a 0.3
iii
Gi 0.3
~ 0.2 ~ 0.2-1---""",-
~
~ 0.1
ft
~ 0.1

a::: 0 0 0.4 0.8 1.2 1.6 2 2.4 2.8 0::: 0 0 0.4 0.8 1.2 1.6 2 2.4 2.8
DISTANCE FROM THE WAlL FACE :m DISTANCE FROM THE WAlL FACE :m

Fig. 7. Predicted reinforcement force distributions: 12 kPa

surcharge, ideal boundary conditions.

The truncated length case involves an idealised reinforcement

force distribution with a uniform force (equal to the maximum
force) along the whole length of the reinforcement. Clearly a
bond length would be required on the end of the reinforcement
layers to generate the reinforcement force if this limiting
state of equilibrium were to be achieved.
The two "bounding" reinforcement force profiles are shown
plotted together in Fig. 7a to d for the four reinforcement
layers in the trial wall and for the case ~ = 50'. The range
between the two force profiles has been shaded in the figure.
The actual distribution of the reinforcement force would be
expected to lie within the shaded zone. '
There are two points which should be brought out:
Firstly, the "bounds" only apply to the distribution
of the reinforcement force behind the most critical
mechanism, the maximum force in each layer is the same
for the two equilibrium states.
Secondly, the shape of the predicted force profile
qhanges with the elevation of the reinforcement layer
in the wall, Fig. 7. The stressed reinforcement
length is greater towards the top of the wall. The
predicted "rate of change" in the reinforcement force
(ie the force transfer with the soil) is much steeper
lower in the wall.
210

5.6 Reinforcement elongation

The effect of a bond length on the overall elongation in the
reinforcement is small for the truncated length case for wide
width polymer reinforcement materials. These reinforcements
typically have a high bond capacity compared with the working
force magni tude. Tae bond length is ignored in the
calculation of the reinforcement elongations in the truncated
length case. This is consistent because the truncated length
case aims to represent the minimum likely reinforcement
length, and hence the minimum likely reinforcement elongation.

Table 5. Calculation of reinforcement elongations:

12 kPa surcharge for 100 hours duration.

Reinforcement Elevation z {)K P ()

:m H HP :kN :mm

Layer 4 2.50 0.17 0.36 2.30 4.0

Layer 3 1. 75 0.42 0.33 3.38 5.2
Layer 2 1. 00 0.67 0.26 4.64 5.8
Layer 1 0.25 0.92 0.13 4.88 3.1
Notes: K= 625 kNlm; H- 3000mm.

The distribution of elongation in the reinforcement layers can

be derived directly from the non-dimensional charts given in
the companion paper. The elongation is simply scaled from the
chart at the elevation of the reinforcement layer in the wall.
The magnitude of the strain is calculated by using the maximum
reinforcement force and the reinforcement stiffness for the
particular wall.
The procedure is illustrated in Fig. 8a for the truncated
length case and a mobilised soil strength if> = SO·. The
calculated reinforcement elongations are shown in Fig. 8b as
the implied horizontal movement at the face , measured in mm.
The numbers are entered in Table 5. The procedure may be
repeated for the ideal length case, and for the other
mobilised soil strengths, using the appropriate charts in the
companion paper.
The results for both equilibrium cases are shown plotted
together for the three soil strengths in Fig. 9. The actual
distribution of reinforcement elongation is expected to lie
between the two predictions, and this range has been shaded.
One point to notice is the relatively small difference between
the predicted reinforcement elongation for the two equilibrium
states at about half the wall height. The main difference
between the two equilibrium states occurs at the top and
 211

bottom of the wall. The difference between the two "bounding"

predictions also indicates the order of accuracy which might
be anticipated from the prediction.

"lRUNCATED REINFORCEIIENT I.£NGIH CASE

0
~
i I Ii
• 11IUIICATED I.iIIGIH

/ ~
0.2

I .
l!; 04

/ !
1 06
~
/
.

~ 0.8 ~
...- ~
1
0 0.2
OUTWARD MOVEMENT
(~ 0.4
0
0 2 •
REINFORCEMENT ELONGATION :mm
10

(A) (8)

Fig. 8. (a) Use of a non dimensional chart (b) to calculate

reinforcement elongations.
Another important point to remember concerning these results
is that the reinforcement elongations do not represent the
total outward movement at the face. The influ.ence of the
"moving datum" due to incremental construction has to be
added.

!'i
3
... 411. lRUNCATED
411. IDEAl.

0::
w
~
(!) 2
z
5
z
::J
m

~
$!
z
0
~
~
0
0 2 4 6 8 10
REINFORCEMENT ELONGATION :mm

Fig. 9. Range of reinforcement elongations: 12 kPa surcharge

and ideal boundary conditions.
212

ADDITIONAL MOVEMENT FROM INCREMENTAL CONSTRUCTION

6.1 Introduction
The additional movement from incremental construction
represents a "moving datum" for the reinforcement layers
because they are attached to panels which are aligned over the
previous layer of panels which have already moved outwards.
At any elevation where there is reinforcement in the wall, the
net surveyed outward movement of the face from the vertical
line through the original position of the lowest panels is
made up from (1) the elongation in the reinforcement layer and
(2) the initial position of the unstressed (or lightly
prestressed) reinforcement layer from the vertical line.
6.2 Panel geometry and movement
The trial incremental wall is built from four 0.75 m high
panels each with one reinforcement layer attached, which means
that the incremental displacement will be the saIl1e for each
construction lift. The reinforcement is attached to the panel
at the lower third point, and the layers of panels are not
connected top to bottom. This gives each panel the maximum
freedom of movement.
The actual panel movement in the trial wall is unlikely to be
pure translation because the net soil stress on the panel will
act higher than the lower third position of the equilibrating
reinforcement force. The panel movement is also unlikely to
be pure rotation which would occur if the bottom of the panel
were fixed to the top of the panels in the layer below. Both
these extreme cases are calculated, so that the actual net
movement would be expected to lie somewhere in the range.
6.3 Soil strength and compaction load
Incremental construction movement occurs during ini tial
filling and compaction. The deformations in the soil and the
force in the reinforcement are relatively small. Therefore
the appropriate mobilised angle of friction for the soil
should approximately represent "at rest" conditions or a
little higher. For the RMC sand this would be equivalent to
¢i ~ 35°. A range of values rjJ = 30°.35° & 40° are used below for
comparison.
The other parameter which must be evaluated is the uniform
surcharge equivalent to the compaction and construction
loading. For the trial wall the dead weight of the vibrating
plate compactor is 1 kN/m2. The static equivalent to the
dynamic load depends on the frequency and amplitude of
vibra'tion. Construction loading also comes from people and
equipment moving the fill, and any temporary overfilling. The
 213

choice of equivalent load to represent all of these for the

prediction must be somewhat speculative. A value 4 kN/m2 was
selected (equivalent to 0.2 m of temporary overfilling).

6.4 Incremental construction movement

The values of the parameters for the trial wall ( H,.,= O. 7Sm,
q , = 4kNlm2,n= l , y= l8kNlm3 , K= 62SkNlm) can be substituted
directly into the equations in the companion paper, section
5.2, to determine the incremental displacement in each layer
of construction.

;3 ;3
g g

~ ~
"~ 2 "~z 2
::J ::J
01 01

I
""-
~
..30_
40_
1lWOIIJJING PNtIl.S

.. 40_
~
z z
~ + ... -
~ + ... -
~ ~ 00
2 4 6 8 10 12 14 4 8 12 16 20 24
OUlWARD MOVEMENT :mm OUlWARD MOVEMENT :mm

(a) (b)

Fig. 10. Incremental construction induced movement: ( a)

translating and (b) rotating panels.

The results are plotted in Fig. lOa and b as the cumulative

movement due to incremental construction at each panel level .
Both the pure translation and the pure rotation results are
given, the actual panel movement is likely to be between these
two extremes. The data points in the figures show the initial
posi tion for each panel with respect to the vertical line
through the initial position of the toe of the wall. The
calculated outward movement of the top of each panel is about
0.2% to 0.5% of the panel height.

6.5 OVerall movement of the wall face

The overall outward movement at the wall face can now be
determined. In the trial wall this movement is due to the
elongation in the reinforcement (calculated in section 5.6)
and the moving datum caused by incremental construction
(calculated above). There is no movement due to foundation
214

settlement because of the rigid base, and negligible movement

from the unreinforced fill behind the reinforced zone because
of the long reinforcement length.

3,---------------____________________,

NO CONSIRUCTION MOVEMENT
• .50. TRUHCA.TED
+ 50. IDEAl.
WITH CONSIRUCTION 1.t0VEM
~ 50. lRUNCATED

2 6 8 10 12 14 16 18 20
FACE MOVEMENT :mm

Fig. 11. Total outward movement at the face due to

reinforcement elongation and incremental construction: 12 kPa
surcharge and ¢ = 50'.

The predicted total outward movement at the wall face with

ideal boundary conditions is shown in Fig. 11 for both
postulated equilibrium states, and for the case of the final
mobilised soil strength ¢=50' and with the average translation-
al and rotational incremental panel displacement calculated
with a mobilised soil strength ¢=35'. These two components of
movement were recorded in Figs. 9 and 10. Results for other
combinations can be derived directly from these.

SIDE WALL FRICTION IN THE TRIAL WALL

7.1 Boundary conditions in the trial wall

The influence of boundary conditions were reviewed in the
companion paper, section 9. The calculations for the
incremental wall have been based on assumed ideal boundary
conditions. That is the face is assumed to have no stiffness
and to ideally provide the required horizontal stress on the
boundary of the soil exactly balanced by the connecting
reinforcement force. There is assumed to be no net vertical
and horizontal force in the face transmitted through the
 215

connection between the face and the rigid base. Finally, the
rough strong base to the wall is assumed not to influence the
force in the lower reinforcement layers.
In the incremental wall the facing panels do act freely from
one another which is probably as close to the ideal
assumptions as possible. Thus the interaction of the facing
and the base should only influence the force in the lowest
reinforcement layer. The maximum possible reduction in the
reinforcement force in the lower layer due to the rough base
was estimated to be about 20% (section 4.1).
7.2 Side wall friction
The problem always encountered with laboratory and field
trials is at the side boundaries to the structure. Side wall
friction can provide significant stabilising forces in the
soil. Attention was paid to the possible influence of the
side walls for the RMC trial and a measured value of side wall
friction I/J,w = 20' is reported from a large scale shear box test
on the prepared boundary (bulletin 1).
Bransby and Smith (1975) made detailed experimental observa-
tions and presented results from a numerical analysis for the
influence of side wall friction on model retaining wall tests.
An analysis is derived below to evaluate the influence of side
wall friction in the RMC trial wall. A simple closed form
analysis is presented which slightly underpredicts the results
of the numerical solution by Bransby and Smith (1975).
7.3 Side wall friction: analysis for self weight loading
If there is side wall friction then the shear stresses
generated between the side wall and the deforming soil will
act to resist the deformation in the soil. Because there are
additional resisting forces, the side wall forces reduce the
wall pressure required to support the soil, or reduce the
required reinforcement forces in a reinforced soil wall.
Bransby and Smith (1975) show that the side' wall shear
stresses change both the failure mechanisms and the stresses
in the soil. For a simple analysis an assumption can be made
that the critical failure mechanism in the soil is not
changed, and that the side wall friction only acts to alter
the overall force equilibrium. The calculation below is for
the triangular wedge of soil between the most critical plane
and the wall face, the equilibrium of which determines the
maximum reinforcement force.
The most critical plane and the equilibrium force polygon for
the soil wedge are shown in Fig. 12a and b . The required
horizontal force for equilibrium Po is provided by the
reinforcement layers. A side wall force p. acting in the
opposite direction to the movement of the soil wedge changes
the force equilibrium as shown in Fig. 12c and d. For a
simple analysis the soil movement can be assumed to be
parallel to the most critical mechanism.
216

(a) (b)

(e) (dJ (e)

Fig. 12. The influence of side wall forces on self weight

equilibrium in a retaining wall.

The side wall force is derived from the normal stress and the
frictional shearing resistance between the sand and the side
wall. The force is found from summing the shearing stress on
the side wall over the cross sectional area of the critical
wedge, as shown in Fig. 12e.
The intermediate stress a 2 has been observed to remain a
constant function of the mean stress s for plane strain
loading,
(6)

and Stroud (1971), for example, measured K 2 = 9.74 for sand in

plane strain.
The mean stress Sz at any depth in the critical wedge is a
function of the vertical stress in the soil and the active
earth pressure coefficient

(7)

The gross side wall force p ... on a wall of height H can be

found from integration over the cross section of the critical
wedge to give

(8)
 217

where ~,w is the mobilised angle of friction on the side wall,

and w is the weight of a unit thickness of the critical wedge.
The net side wall force per unit thickness of soil P, is a
function of the height to width ratio for the wall H/w. From
eqn. (8) and remembering there are two sides,
Ps K2(1+KaJ 2H
W = 2 tan¢sw3w (9)

The reduced active earth pressure coefficient due to side wall

friction Ka' is a direct function of the net side wall force.
The wedge equilibrium forces are shown in Figs. 12b and d.
The active earth pressure coefficient is

Pa = Ij( . (10)
W V1\.a

and the apparent earth pressure coefficient is defined in

terms of the total soil weight so that
P ,
~ =~Ka' (11)

Resolving the net side wall force into horizontal and vertical
components, and by symmetry (Fig. 12d and eqn. (10))
( P a ' + P sh J = fK: (12)
(w -?v )
which gives a direct expression for the reduced active earth
pressure coefficient, from equations (11) and (12),

(13)

The components of the net side wall force are, from Fig. 12,
P sh = P s cos ( 45 + ¢ / 2)

P sv = P s sin (45 + ¢ / 2)

Comparison with Bransby and Smith (197~)

Two numerical results for the influence of side wall friction
on smooth retaining walls, the case above, are given by
Bransby and Smith. They give the percentage reduction in the
active earth pressure coefficient for soil with an angle of
friction if! = 35° and 50°, for a wall geometry H /w = 2 and with
side wall friction ¢"",= 5.71° (ie. f.1= 0.1). Their calculation
is for an intermediate stress ratio K2 = 0.37. The results from
the Bransby and Smith numerical analysis are compared with the
simple closed form analysis in Table 6.
218

Table 6. Percentage reduction in the active earth

pressure coefficient.

Angle ot" friction: 35· so·

Bransby and Smith -11 % -13 %
Eqns: ( 9 ) & (13 ) -11 % -10 %
Note: Case with r/J~- 5.71' ; Hlw= 2.

The two analyses compare well although the simple closed form
analysis giving a slightly lower reduction in the active earth
pressure, coefficient. One obvious conservatism in the simple
analysis for self weight loading is that it ignores the slight
reduction in the vertical stress in the soil due to !' arching"
between the side walls.

Application to the RMC trial wall

The RMC trial wall has a height to width ratio Hlw= 1.25, and
Bransby and Smith's results indicate a reduction in the active
earth pressure coefficient of approximately 7% for "full
friction". The simple analysis gives a similar result, a 6.5%
reduction for r/J = 50'.
The numerical results presented by Bransby and Smith (1975)
are not directly applicable to the RMC trial wall, however,
because of the two following points.
(1) "Full friction" in Bransby and Smith is an angle of
side wall friction r/J..,= 5.71' for sand against glass.
When the measured value of side wall friction for
the RMC boundaries is allowed ~..,= 20' the percentage
reduction in the earth pressure coefficient changes
from 6.5% to 22.5%.
(2) Bransby and Smith appear to have estimated the
intermediate principal stress with a coefficient
K 2 = 0.37. The data for sand i{l plane strain
generally indicates a higher value K 2 '" 0.65 to 0.75.
The relationship between the parameter b

and K2 for sand with a mobilised angle of friction r/J

can be deduced from a Mohr circle
K 2- 1
b=O.5+--
2 sin cp

Values of K 2 = 0.65 to 0.75 for sand with r/J= 40' to SO'

correspond with values b = 0.25 to 0.35 which is the
commonly observed range for plane strain tests on
sand.
 219

Increasing the intermediate stress to K2 =O.7 in the

closed form analysis increases the predicted
influence of the side wall friction further from
22.5% ((1) above) to 40%.
Apparent active earth pressure coefficient
The result is that the influence of side wall friction on the
self weight loading of the sand in the RMC trial wall is
likely to be significant. If the measured side wall friction
were mobilised in the trial wall, the apparent active' earth
pressure coefficient would be 40% less than the expected earth
pressure coefficient without side wall friction,
Ka'= O. 6K a

7.4 Side wall friction: analysis for surcharge loading

Without side wall friction, a uniform surcharge loading
increases the vertical effective stress in the soil equally at
every depth behind the wall. Side wall friction reduces the
net vertical stress in the soil due to the surcharge. This is
in addition to reducing the active earth pressure coefficient,
as described above.
To allow for this effect in the RMC trial, an analysis is
required for the reduction in the net vertical surcharge
stress in the soil at any depth due to the side wall friction.
Once again, a straightforward closed form analysis is
presented.

qz/qo
0+0_ _...1.-_--11 1

z/H

(a ) ( b)

Fig. 13. (a) Illustration of the soil stresses due to

surcharge loading and (b) the reduction due to side wall
friction.

A vertical wall supporting a uniform surface surcharge qo is

shown in Fig. 13a. At a depth z below the wall crest the
vertical stress due to the surcharge loading is qz. Following
the same analysis as before, the intermediate stress due to
the surcharge load is
220

(14 )

The vertical component of the side wall shear stress acts

cumulatively to reduce the vertical stress in the soil.
Remembering there are two sides to the trial wall of width w,
the rate of change of the net vertical loading can be
expressed as

(15 )

which has the form

The solution to equation (15) is

(16)

where the constant C is defined above.

Results for surcharge loading
The net vertical stress due to the surcharge loading can be
represented by the ratio q./q.. The RMC trial wall is 2.4 m
wide. With an angle of friction in the soil ¢!= so· and using
the measured angle of friction for the side wall ¢!.w = 20· the
constant c = 0.11, assuming a .value K 2 = 0.7 for the intermediate
stress. The resulting variation in the net vertical stress
due to surcharge loading with depth is shown in Fig. 13b.
These results indicate that the vertical surcharge loading
would be reduced by approximately 30% at the bottom of the RMC
trial wall due to the side wall friction.
7.5 Implications for the RMC trial wall
The analysis presented above indicates that side wall friction
is likely to have an important influence on the RMC trial
wall.
For self weight loading in the wall the side wall friction
effectively provides about 40% of the required horizontal
stresses for equilibrium, or, in other words, results in an
apparent active earth pressure coefficient Ka'= O.6Ka (section
7.3). This reduction in the required horizontal stresses for
equilibrium is shown in Fig. 14a.
The side friction also leads to "arching" for any vertical
surcharge applied to the RMC trial wall. At a depth 3 m below
the c"rest of the wall the side walls support about 30% of the
applied surface surcharge load, (section 7.4). This was shown
in Fig. 13.
 221

(a) Self weight loading (b) Surcharge loading

Fig. 14. Illustration of the reduction in the required soil
stresses due to side wall friction in the RMC trial wall.

The reduced vertical stress also requires proportionally less

horizontal stress to maintain equilibrium in the soil, because
of the side walls. The apparent active earth pressure
coefficient is the same as that calculated for the self weight
loading, 40% less than the normal value, giving the required
horizontal stress for equilibrium shown in Fig. 14b.

PREDICTION ALLOWING FOR SIDE WALL FRICTION

12 kPa applied for 100 hours
50 kPa surcharge applied for 1000 hours

8.1 Introduction
A prediction for the RMC trial wall allowing for the influence
of the side wall friction can now be completed. The
prediction is for the expected plane strain mobilised angle of
friction in the soil ~= 50' (section 2.4), and for the measured
angle of side wall friction ~~= 20'. The prediction is for the
central loading case (12 kPa/100 hours) and for the end of the
test (50 kPa/1000 hours).
For both loading cases the difference between the prediction
with and without the side wall friction is indicated.
An implicit assumption in the analysis for side wall friction
is that it only reduces the magnitude of the required
reinforcement forces for equilibrium and does not change the
location in the soil where these forces are required. This
assumption is likely to be conservative. The side wall
friction is likely to reduce the overall elongation in the.
reinforcement layers more than predicted.
222

8.2 Force in the reinforcement

Self weight loading
The distribution of the gross required horizontal force for
equilibrium to the individual reinforcement layers was
described ln section 4.1. Side wall friction reduces the
gross required force but not the distribution of the force.
The percentage of the reduced gross required force due to self
weight carried by each of the four reinforcement layers is
given in Table 7, together with the force magnitudes. The
force in each layer can be compared with the results in
Table 4.

Table 7. Maximum reinforcement force:

12 kPa surcharge and side wall friction.
Loading Self weight Surcharge Self weight Surcharge Total
:kN :kN :kN
Layer 4 9 % 33 % 0.57 0.78 1. 35
Layer 3 21 % 26 % 1. 33 0.61 1. 94
Layer 2 33 % 24 % 2.08 0.56 2.65
Layer 1 37 % 18 % 2.34 0.43 2.77

Total 100 % 100 % 6.32 2 . 38 8.70

Surcharge loading
The same form of presentation can be used for the surcharge
loading. In this case the gross required force is reduced by
more than the 40% due to the apparent earth pressure
coefficient. The reduction of stress in the soil due to the
"arching" over the side walls (section 7.4) gives an
-additional overall reduction in the gross required force of
15%. The resulting reduction in the gross required horizontal
force to support the surcharge load is 49%.
The "arching" of the surcharge loading alters the distribution
of the gross required force between the reinforcement layers.
The distribution between the reinforcement layers of the
reduced gross required force due to surcharge loading is given
in Table 7. The magnitude of the resulting forces in the
reinforcement layers for 12 kPa surcharge loading are also
given. The values can be compared with the results in
Table 4.
Total reinforcement force
The resulting maximum force in the four reinforcement layers
for the central loading case and at the end of the test are
shown plotted in Fig. 15a and b, where they are compared with
the values without side wall friction.
 223

Maximum Struln Waxlmum struln

3
005 01 ·905 .Ql .'!I5
12 W'o at 100 houno Ii 3
Ii • SIIOO1H sa: ..w
~2.6 ~ 2.6
':52.2 \ l"".
+ SIDE WAlL. RICIXIN
':5 2.2 1
~ 1.B \ ~ ~ 1
i1.~
z ~ 1.8
~
.""[\
:::J :::J
1.4

i
II>

\ 1

1 \ ~
liO !cPa 1000 HOURS
0 .6 ~0.6 • IIIOCJrtt SIIE 1N.lS

0.2 0 2 4
\ 6 B 10
iii
:>: 0.2
o
.. SllEIIIU...:::noN

2 4 6 8 10
MAXIMUM REINFORCEMENT FORCE :kN/m MAXIMUM REINFORCEMENT FORCE :kN/m

Fig. 15. Maximum reinforcement force with and without side

wall friction (a) 12 kPa surcharge and (b) 50 kPa surcharge.

8.3 Strain in the reinforcement

The maximum strain in the reinforcement depends on the maximum
force and the reinforcement stiffness (section 3.1). Using
the appropriate value of stiffness for the two loading cases
( K 100 = 625kN 1m and K 1000 = 550kN 1m ) gives the set of maximum
reinforcement strains shown on the top scale in Fig. 15.

Table 8. Calculation of reinforcement. elongations;

12 kPa for 100 hours and side wall friction.

Reinforcement Elevation z oK P 0
:m H HP :kN :mm

Layer 4 2.50 0.17 0.36 1. 35 2.3

Layer 3 1. 75 0.42 0.33 1. 94 3.0
Layer 2 1. 00 0.67 0.26 2.65 3.3
Layer 1 0.25 0.92 0.13 2.77 1.7
Notes: K= 625kNlm; H= 3000mm.

8.4 Reinforcement elongation

The procedure for evaluating the total elongation in each
reinforcement layer is the same as before (section 5.6). In
the case of side wall friction the reinforcement force is
lower, which reduces the overall elongation.
224

The calculation for the central loading case and the truncated
length equilibrium case, and allowing for side wall friction,
is set out in Table B. First the depth of each reinforcement
layer beneath the wall crest is determined and the non
dimensional outward movement read from the relevant chart
(Fig. 8a for this case, but see the companion paper for the
other charts). The maximum force in each layer is recorded in
Table 7, and the reinforcement stiffness and the wall height
are the same for all the layers. The magnitude of the outward
movement can be directly evaluated, Table B.
The predicted set of reinforcement elongations for the two
equilibrium states of ideal length and truncated length are
shown in Figs. 16 and 17. The difference due to the side wall
friction is indicated for the two loading cases.

3
NO SIDE FRICTION
~ • lRUNCAlED SIoAOOTH
0:: ,. IDE'AL SIoAOOTH
W
WITH SIDE FRICTION
~ <> lRUNCAlED ROUGH
t. IDE'AL ROUGH
C,!)
z 2
is MEASURED
z l( PANELS
::J v WIRES
m
w
e;
~
z
0
~
~
w
0
0 2 468 10
TOTAL REINFORCEMENT ELONGATION :mm

Fig. 16. Predicted reinforcement elongations for the 12 kPa

surcharge loading, compared with the measured values.

8.5 Incremental construction movement

The height to width ratio for each incremental lift is high
enough to allow the influence of the side wall friction to be
ignored. The previously calculated movements caused by
incremental construction can be used (section 6.4).
 225

3~------------------------------------~

IZJ WITH SIDE WAll. FRIGnO

- MEASURED
0+-~~~~~~~~~~~=T=r~T=~
o 2 4 6 8 10 12 14 16 18 20
TOTAL REINFORCEMENT ELONGATION :mm

Fig. 17. Predicted reinforcement elongations for the 50 kPa

surcharge loading, compared with the measured values.

PREDICTED BEHAVIOUR FOR THE INCREMENTAL TRIAL WALL

9.1 Introduction
The analysis for walls involves two idealised equilibrium
states which are expected to bound the actual equilibrium in
the reinforced soil. The data from the measured behaviour of
the trial walls have been added to the figures in this and the
previous section.
9.2 Deformations
The total deformation at the wall face is made up from the
reinforcement elongations and the "moving datum" caused by
incremental construction.
226

Reinforcement elongation
The predicted reinforcement elongations for the central
loading case are shown in Fig. 16. The results for the ideal
length case are also shown, as are the maximum likely
elongations that were predicted assuming no side wall
friction.
The predicted reinforcement elongations at the end of the test
are shown in Fig. 17. Again, the companion ideal length case
is also given, as are the maximum likely elongations
calculated ignoring the side wall friction.
Incremental construction movement
Because the facing panels can either translate or rotate there
is a range of possible movement due to incremental
construction. The predicted range is shown in Fig. 18, based
on a mobilised angle of friction if> = 35' consistent with the
initial filling and compaction of the sand (section 6.3).
3~-----;~~~~------------------,

MEASURED
.. PAt-JEL TRANSLATION
Q PANEL ROTATION

0~---r---,----,----,---,----'---'----1
o 10 20 30 40
INCREMENTAL CONSTRUCTION MOVEMENT :mm

Fig. 18. Predicted incremental construction movement, com-

pared with the measured values.

9.3 Reinforcement force and strain distributions

In th.e zone between the most critical plane and the wall face
the predicted reinforcement force is uniform, and the
magnitude is the same for both the ideal length and the
truncated length equilibrium states. The variation of the
reinforcement force behind the most critical plane does depend
on the the equilibrium case, and the variation for both cases
is shown.
 227

The predictions are presented in terms of the tensile

reinforcement strain as this is what is actually measured.
The corresponding force is governed by the appropriate
reinforcement stiffness (section 3.1). As usual, the
prediction is for the case with side wall friction.

PREDICTED AND MEASURED REINFORCEMENT STRAIN PROFILES

12 kPo SURCHARGE FOR 100 HOURS
0.6r--------------, 0.6
z
~ O~ LAYER 1
~ 0.5
LAYER 3

!z 0.4 T."~. !z
w
0.4

~ O~ Ci
(,)
0.3
..-
~ 0.2 VI '"~ 0.2

1\j 0.1 V z 0.1

iii
0::: 0
o
......
0.4 0.8 1.2 1.6 2 2.4 2.8
'" 00
DISTANCE FROM THE WALL FACE :m DISTANCE FROM THE WALL FACE :m

0.6 0.6
z z
LAYER 2 LAYER 4
~ 0.5
0.4
~
!z
0.5
0.4
!z
w
Ci 0.3 ill 0.3
(,)
ii 0.21-7"<,-._-....
'"z~ 0.2
~
Z 0.1
0.1
iii
ill
~

'" 00 0.4 0.8 1.2 1.6 2 2.4 2.8 00 0.4 0.8

DISTANCE FROM THE WALL FACE :m DISTANCE FROM THE WALL FACE :m

Note. Reinforcement strain is in units of percent (%).

Fig. 19. Predicted reinforcement strain profiles for the 12

kPa surcharge loading, compared with the measured values.

The predicted reinforcement strain in the uniform zone at the

front of the wall, and the predicted variation of strain
behind this zone for the ideal length and the truncated length
cases, are shown in Fig. 19. These predictions are for the
central loading case and for the four reinforcement layers.
The reinforcement stiffness K 100 = 625 kN 1m •
The corresponding predictions at the end of the test are shown
in Fig. 20. The reinforcement stiffness K 1000 = 550 kN 1m.
9.4 Stress on the base
Because of the side wall friction, the predicted vertical
stress on the base of the wall is less than the sum of the
overburden and the surcharge stress. The equilibrium in the
reinforced soil has a uniform vertical stress on the lower
boundary.
228

In the simple analysis for side wall friction the reduction in

the self weight vertical stresses was ignored, "arching" was
only calculated for the surcharge loading. The reduced
vertical stress at the base of the wall due to surcharge
loading by 30%, (section 7.5). Thus the total predicted
uniform vertical stress for the two loading cases is 62 kPa
and 89 kPa.

PREDICTED AND MEASURED REINFORCEMENT STRAIN PROFILES

50 kPa SURCHARGE FOR 1000 HOURS

!:/~
z
LAYER 1 LAYER 3
~ 0.8

!z
w
0.6

2i
- _.
:::!
() 0.4
w
() 0.4
0:: 0::
~
z 0.2 ~ 0.2
Z
iil
0:: V·,
O~~~~~~~~~~~~
iil
0::
O~~~~~~~~~~~~
o 0.4 0.8 1.2 1.6 2 2.4 2.8 o 0.4 0.8 1.2 1.6 2 2.4 2.B

,
DISTANCE FROM THE WAlL FACE :m DISTANCE FROM THE WALL FACE :m

z z
LAYER 2 LAYER 4
~ 0.8 ~ 0.8
+--f-J~=
!z
w
0.6 !z
w
0.6
:::!
w
()
0::
0.4 ~
0::
0.4

~ 0.2 ~
Z 0.2
Z
iil iil
0:: tl!
O~~~~~~~~~~~~
o 0.4 0.8 1.2 1.6 2 2.4 2.8 00 0.4 O.B 1.2 1.6 2 2.4 2.B
DISTANCE FROM THE WALL FACE :m DISTANCE FROM THE WALL FACE :m

Note. Reinforcement strain is in units of percent (%).

Fig. 20. Predicted reinforcement strain profiles for the 50

kPa surcharge loading, compared with the measured values.
 229

THE PROPPED WALL TEST

10.1 Magnitude of face movement

Attention in this paper has been given to the incremental
wall, the most usual form of reinforced soil wall
construction. The analysis and prediction for the incremental
wall has provided much material, and space does not permit a
similar analysis for the propped wall.
However the RMC trial walls do provide an outstanding example
of the contrast between an incremental and a propped wall. In
an incrementally constructed wall there is a "moving datum"
which results in deformation at the face additional to the
deformation caused by the reinforcement elongation. This
additional movement does not occur in the propped wall because
the propping supports the face so that the initially
"unstressed" reinforcement layers all align on a vertical line
through the toe (or at the angle of the propped face).
In the RMC trial the predicted incremental movement due to
construction is about a factor of four greater than the
predicted movement due to the reinforcement elongation (Figs.
16 and 18). This means that the movement in the propped wall
should be only about 20% of the movement in the incremental
wall - a very significant difference.

§ :3

~
(!)
z 2 PROPPED WALL
az PREDICTED
::J
CD & 1lDL .......

-
~
• lRUNCAlED IIOUGH
MEASURED
~ T
Z
0
~
~ 0
0 2 4 6 8 10
TOTAL REINFORCEMENT ELONGATION :mm

Fig. 21. Predicted face movement for the propped wall under
12 kPa surcharge, compared with the measured movement.

The predicted face movements for the propped wall at the end
of the 12 kPa surcharge loading, and at the end of the test,
are shown in Figs. 21 and 22. The predicted movement is due
only to the reinforcement elongation and is exactly equal to
the equivalent prediction for the incremental wall. The
prediction allows for the measured side wall friction.
230

10.2 Influence of the stiff face boundary

As discussed in the, companion paper, one possible detrimental
consequence of propped wall construction with a continuous and
stiff face is that the face deformation is constrained to
remain approximately linear. The analysis that has been
presented indicates that this type of face movement only
occurs freely for the ideal length equilibrium state combined
with ideally spaced reinforcement. When more uniform
reinforcement spacing is used the face movement towards the
top of the wall is reduced. The results in Fig. 21 are a
typical example of this.
The consequence is that close to the top of a propped wall the
stiff face might tend to "pull the reinforcement outwards" to
meet the required boundary displacement, thereby causing
higher than expected reinforcement force close to the face at
the top of the wall.

3
Ii

~
"isz 2 PROPPED WAIl.
PRf.DICIe)

·
•
~ IIE'II._
11IIJNCATED _
lD

I
Z
0
•
II£ASURED
WIllES

~
~
2 4 6 8 10 12 14 16 18 20
TOTAL REINFORCEMENT ELONGATION :mm

Fig. 22. Predicted face movement for the propped wall under
50 kPa surcharge, compared with the measur~d movement.

By exact analogy, towards the mid-height of the wall the

outward movement would be reduced by the face boundary
condition thereby locally reducing the reinforcement force
close to the face. The discrepancy between the predicted
movement with an ideal face boundary and the linear movement
with a stiff continuous facing is illustrated in Fig. 21 for
the RMC propped wall.
The conclusion to be drawn is that this phenomenon should be
considered in the design of propped walls to ensure that local
over-'stressing at the reinforcement face connection towards
the top of the wall does not occur.
 231

STRAIN COMPATIBILITY

11.1 Introduction
The aim of this section is to relate the measured soil and
reinforcement properties on a compatibility curve for the
incremental wall (companion paper, section 6.2). Only the
overall compatibility for the whole wall will be considered -
a more detailed analysis would consider equilibrium in each
layer. The influence of the measured side wall friction is
included in the analysis.

11.2 Material properties

The soil is initially at a low stress ratio on filling the
wall, corresponding to approximately K. conditions. The
initial mobilised angle of friction 9= 35· was chosen (section
6.3). This is somewhat less than the critical state angle o~
friction.
The simple model for sand involves elastic shear to the (peak)
mobilised stress ratio, followed by perfectly plastic shear
with a constant angle of dilation. The predictions were based
on a mobilised angle of friction 9 = so· which has a
corresponding angle of dilation 0/= IS· (section 2.4).
The shear modulus for the sand deforming from the initial to
the final stress ratio was estimated as (G/s)~ 23. The initial
stiffness was about twice as high giving the range (G/s)~ 20 to
SO (section 2.5).

11.3 Reinforcement properties

The load extension properties are assumed to be linear and the
reinforcement stiffness varies with time under load. The
reinforcement stiffness values used for the prediction are
K 100 = 62S kN 1m and K 1000 = SSO kN 1m (section 3.1).

11.4 Compatibility curve

The compatibility curve depicts the relationship between the
required force for equilibrium in the structure and the
available force from the reinforcement. The link is the
tensile strain in the soil in the direction of the
reinforcement. The overall compatibili ty curve considers
equilibrium along the most critical plane. Because the
principal stresses are vertical and horizontal on this plane,
the relevant tensile strain is the minor principal strain for
vertical walls.
The simple compatibility curve is drawn ignoring the
construc'tion loading sequence (companion paper). Drawn this
way it gives a much more clearly defined equilibrium point.
232

The self weight and surcharge loading are used to determine

the required force. The required force depends directly on
the shearing resistance in the soil. The link between the
shearing resistance and the tensile strain in the soil for the
trial wall is predicted from the simple soil model and the
analysis of the direct shear test results.
The available force depends on the magnitude of the tensile
strain, the stiffness of the reinforcement and the number of
reinforcement layers.

Central loading case

The compatibility curve for the central loading case (12 kPa
and 100 hours) is shown in Fig. 23. The variation· of the
required force is (almost) bilinear reflecting the elastic
plastic model for the sand. The measured range of the soil
shear modulus is also shown.

E
"-
z 24
-!!
w 20
tJ

'"~
16
~
~ 12
~
0
~ 8
fil
'"5 4
fil'
'" 0
0 0.002 0.004 0.006 0.008
PRINCIPAL TENSILE STRAIN

Fig. 23. Compatibility curve showing the predicted overall

equilibrium point under 12 kPa surcharg~ loading.
The compatibility curve indicates equilibrium in the wall,
where the gross required and available forces are equal,
almost exactly at the point when the soil reaches the plateau
of continued deformation at a constant angle of friction ¢= 50°
The predicted tensile strain at the equilibrium point is
0.4%, which is consistent with the predicted reinforcement
strains in Fig. 19.
The compatibility curve at the end of the test (50 kPa and
1000 hours) is shown in Fig. 24. The equilibrium point occurs
at a larg-er tensile strain 0.8%, which is more than sufficient
for the reinforcement to mobilise the plateau stress ratio.
Again, the magnitude of the equilibrium tensile strain is
consistent with the earlier predicted reinforcement strains,
Fig. 20.
 233

COMMENTS ON THE MEASURED BEHAVIOUR

The measurements made on the trial walls have been plotted

with the predictions in Figs. 16 to 22 for ease of comparison.

12.1 Implications from "back analysis"

Leroueil and Tavenas (1981), in a review of the developments
that can be achieved in geotechnical engineering from field
observation, have highlighted the possible pitfalls of drawing
conclusions from back analyses. One major pitfall is from
" .. the common practice to obtain a good fit only
for one or two parameters of the field
behaviour: .. evaluating settlements but not
lateral displacements .. construction pore pres-
sures but not deformations .. strength parameters
but not stress strain response up to failure."
" .. the most serious shortcoming (of the above
methodology) is that basic soil mechanics
principles are frequently not satisfied by the
results: .. many cases are reported of 'correct'
settlements computed together with erroneou~
pore pressures and effective stresses."

<40
E
~ 35
w
0 30
D:
f2 25
~
~ 20
~ 15
0
~
10
iii
~ 5
~
0.002 0.004 0.006 0.008 0.01 0.012
PRINCIPAL TENSILE STRAIN

Fig. 24. Compatibility curve showing the predicted overall

equilibrium point under 50 kPa surcharge loading.

In their guidelines for the proper use of back analysis,

Leroueil and Tavenas emphasise that "soil mechanics principles
must apply to the results of back analyses ... all interrelated
parameters must be back analysed with equal success .. ",
adding, " .. a back analysis successful on only 50% of the
parameters is nothing but 100% wrong." They go on to say
234

that " .. time is a systematic parameter in all geotechnical

problems and the validity of back analyses should be checked at
various stages of the process under consideration".
Against these criteria, the required predictions and the
measurements recorded for the RMC trial walls present a fairly
rigourous test of any proposed analysis. Except for vertical
settlements, every aspect of the reinforced soil wall behaviour
was requested to be predicted at different times through the
loading history of the wall.
12.2 Observation on the predictions
The agreement between the predictions from the analysis which
has been presented and the measured behaviour plotted in Figs.
16 to 24 is satisfactory in two respects. (1) The predicted
pattern of behaviour as well as the magnitude is similar to the
measurements. (2) The above observation applies equally to
the two loading cases. Also important is that the analYSis
allows for the measured material properties and boundary condi-
tions, and provides a link between the measurements in the two
separate trial walls with a different face and construction
boundary condition.
12.3 Observation on the measurements

An excellent feature of the RMC trial walls was that several of

the measurements were made by more than one independent means:
for example, the reinforcement elongation was measured'directly
with steel wires, estimated by integrating the measured strain
profile along the reinforcement and interpreted from the move-
ment of the facing panels. This. provides confidence in the
magni tude of the measured parameter but also indicates the
order of accuracy that can be attributed to any measurement.
The trial walls clearly show that there is "scatter" in the
behaviour of a reinforced soil wall introduced by the variable
influences of construction. Therefore an "exact" representa-
tion of the behaviour by a single analysis is perhaps an
unrealistic goal. If this is correct, then an analYSis that
"bounds" the likely behaviour (such as proposed in this paper)
might indeed turn out to be the most satisfactory approach.
12.4 Class A Prediction

The author's class A submission to the NATO prediction exercise

varied in some minor details from the results presented in this
paper but was based on the truncated length equilibrium state
allowing for side friction. In that analYSis a mobilised plane
strain. angle of friction for the sand of <l> = 50 0 and the
measured side wall friction of <l>sw = 20 0 was used. A maximum
likely limit was also predicted with the side wall friction set
to zero.
 235

Acknowledgements
The author would like to thank Dr. R. Kaniraj of the Indian
Insti tute of Technology, who carried out four of the direct
shear tests reported in the paper while at Oxford University
on a British Technical Cooperation Award.
The foresight and hard work of Richard Bathurst, Peter Jarrett
and Alan McGown, and their co-workers, was essential in that
they made the prediction symposium possible, designed the
trials to provide the detailed data of wall performance, and
provided the forum for the exchange of knowledge that seems
certain to advance the state of the art.
References
Bathurst, R.J. & Jarrett, P.M. (1986). Class A prediction
exercise for reinforced earth walls. Bulletin No. 1 for NATO
Advanced Research Workshop, Application of Polymeric Rein-
forcement in Soil Retaining Structures, Royal Military
College, Kingston.
Bransby, P.L. & Smith, I.A.A. (1975). Side friction in model
retaining wall experiments. Journal of Geotechnical Engineer-
ing, ASCE GT7, July, 615-632.
Bolton, M.D. (1986). The strength and dilatancy of sands.
Geotechnique 36, No.1, 65-78.
Cornforth, D.H. (1973). Prediction of drained strength of
sands from relative density measurements. ASTM Spec. Tech.
Publ. 523, 281-303.
Jewell, R.A. (1987). Reinforced soil wall analysis and
behaviour. Proc. NATO Advanced Resear.ch Workshop, Application
of Polymeric Reinforcement in Soil Retaining Structures,
Martinus Nijhoff.
Jewell, R.A. & Wroth, C.P. (1987) Direct shear tests on
reinforced sand. Geotechnique 37, No.1., 53-68.
Koerner, R.M. (1968). The behaviour of cohesicmless soils
formed from various materials. PhD Thesis, Duke University.
Soil Mechanics Series No. 16.
Leroueil, S. & Tavenas, F. (1981). Pitfalls of back
analyses. Proc. 10 Int. Conf. Soil Mech. & Fndn. Engng. ,
Stockholm, Vol 1, 185-190.
Stroud, M.A. (1971) . The behaviour of sand at low stress
level in the simple shear apparatus. PhD Thesis, Uni versi ty
of Cambridge.
Case Histories and Full Scale Trials
 POLYMERICALLY REINFORCED RETAINING WALLS AND SLOPES
IN NORTH AMERICA

M.A. Yako, Student, Purdue University, Lafayette, Indiana,

USA
B.R. Christopher, Principal Engineer, STS Consultants, Ltd.
Northbrook, Illinois, USA

1. INTRODUCTION
This paper presents a summary of reported polymerically
reinforced soil wall and slope projects constructed in North
America. The projects included walls and slopes where the
reinforcement was used to resist lateral earth pressures and
prevent internal failure through the face or toe of the
structure. projects where reinforcement has been used to
provide increased stability against deep seated failure of
embankments were not included. Specific emphasis was placed
on projects that have been instrumented and monitored such
that performance assessments could be made.

The projects were obtained through a review of published case

histories and through interviews with public agencies,
geosynthetic manufacturers, and prominent design engineers.
The projects were reviewed with respect to four categories:

General Information
Design Information
Reinforcement Properties and Design Methodology
Instrumentation

2. PROJECT SUMMARIES
The project summaries for each specific review category are
contained in the attached Tables 1 through 4. Table 1
includes general project information and provides references
where additional information may be obtained. Table 2
provides the design requi rements for the specific projects
including height, surcharge loading, and foundation and
backfill characteristics. Table 3 presents information
pertaining to the actual properties of the reinforcement, and
the design methodology used to determine the reinforcement
requirements. Finally, Table 4 presents project
instrumentation and monitoring information. Unfortunately,
only a few cases were reviewed which contained this extremely
valuable and needed information.
In all, summaries of 54 projects are provided in the tables.
The summaries include 13 projects that were at least

239

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 239-283.
© 1988 by Kluwer Academic Publishers.
240

TABLE 1. GENERAL INFORMATION

CASE PROJECT LOCATION YEAR STRUCTURE HEIGHT LENGTH

No. NAME CONSTRUCTED (m) (m)

II J inols Siskiyou Roadway

River national 1974 widening 3.0 20
Road forest Oregon

Olympic Natl. Shelton TlllIber heul

Forest Wash i ngton 1975 road wa II 0.9 to 6.1 50
Road

Theory & pr inc. WES 09slgn:3.7

ot reinforced Vicksburg 1974-1976 Test wall FaI19d:3.0 4.9
earth Mississippi

Cortright Gifford Roadway

Lava Timber Plnchot Natl. 1979 widening 30
Sale forest WA

New York Col umb i a County sl ide correction

Route 22 NY 1980 In hili 4.9 33 and 46
2 walls

Mt. Baker Mt. Baker roadway wIden i og

6 Nat I. Forest Nat I. Forest 1980 across 5 I t de 1.8
WA area

C...p Hili Wi 11 anette roadway wI den log

Road Natf. Forest 1981 across s 11 de 8.5 88
DR area

State route Mamoth lakes Road wi den i ng

203 CA 1981 2walls 0.9 to 3.7 230

Interstate Glenwood Highway

1-70 Canyon 1982 embankment 4.6 91
test wall CO test wall

------------------------------------------------------------------.::----------------------------
Water down Slope failure
10 CP Rail Ontario 1982 repair 4.6 and 5.5 91

Highway Beaumont Slope failure

11 US 69 TX 1982 rep a ir 6.7

Wauwatosa Reta in i ng
12 Wauwatosa WI 1982 wall 8.2 64

Devil"'s Punch Oev i I Os Punch Lands I ide Bottom 21

13 Bowl Slide Bowl State 1983 stabll i zation 9.1 Top 52
Correction Park OR

Hi Ilcrest San Pablo Lands I ide

14 Road CA 1983 stabi lization 11.9 150
for roadway
 241

FACING RE I NFORCEMENT JMMENTS REFERENCES

TYPE BRAND POLYMER

Wrapped,1:8 Nonwoven F i bertex 1s t fabric

5 lope need I epunched 400 Polypropylene we r I in U.S. 1,2
gun ita

Wrapped,vert. Non ..oven Fibertex Durabi I ity

essl asphalt needJepunched 400 and 600 Polypropylene treat/untreated 1,3 ,4 , 8
emulsion coat fabr ic

Nonwoven Bidim
needlepunched C-28 & C-38 Pol yester

AlcoaT11 4 ply Nylon Neoprene Fai I ure maybe

high strength strip heavy duty coated n:: Ion I arge deform 5, 6
,Janals,vertlcal fabr Ic of ties

Wrapped aspha I t Nonwoven F i bertex

emu I 5 ion need I epunched 200 Po I ypropy I ene
coating

Wrapped f 1 : 3 Nonwoven Bldim Crushed stone

5 lope re i nforce need I epunched C-34 Polyester below 11ft to 7, 8
gun i ta decrease sliding

Wrapped asphalt Woven silt Supac Sawdust

emulsion film 5W Pol ypropy lene back! III
coati ng

Asphalt Woven 51 it Supac Sawdust

emulsion, latex film 5W Po I ypropy I ene backf II I for
paint sl ope stability

Rai froad Tire sidewall withstood

ties wi tie back earthquake of
anchor bar 5.8

Wrapped 1: 5 Nonwoven F I bertex Wall Incorp

slope reinforce needlepunched 200 and 400 Pol ypropy lene tour nonwoven 8,10,11
gun Ita geotextlles

Nonwoven Supac Each in 2

need I epunched 4"" & 6NP Polypropylene weights in 10
test segnents

Nonwoven Trev I ra
needlepunched S1115 & S1127 Pol yester

Nonwoven Typer
heatbonded 3401 & 3601 Polypropylene

Wrapped 1.5: 1 Tensar Pol yethy I ena hi ground waTer

slope Geogrld SR2 HOPE pressure caused 12,13,14
Initial failure

Slope 3: 1 Tensar On site sll1y

Tenser SS2 Geogrld SS2 Pol ypropy lene cl ay excavated 15
on s lope face, , recompacted

Wrapped aspha I t Nonwoven Western

emulsion needlepunched Supac 12NP Pol ypropy lene bu II der 1983
coating

Wrapped 1:6 Tensar Polyethylene First geogrld

slope UV Geogr Id SR2 HOPE retaIning wall 8,16,17
stable grid In U.S.

Wrapped 1:4 Tenser SR2 Polyethylene First use at

slope Geogr l d and SS2 HOPE googr I ds to 18
Polypropylene repair landslide
242

TABLE 1. GENERAL INFORMATION (cont'd)

CASE PROJECT LOCATION YEAR STRUCTURE HEIGHT LENGTH

No. NAME OONSTRUCTED (m) (m)
----------------------------------------------------:-- -------------------':'"~~~--------~

Jewell Jet. Temporary wa II

15 U.S. 26 OR 1983 for stage 4.9 114
construction

Oanapo i nt Reta i ned f II I

16 Oanapa I nt CA 1983 for park Ing 3.0 91
lot

LaHonda Slope failure

17 Route 84 CA 1984 repair 14.0 70

Tarque Verde Tuscon Grade sep. for

18 Road AZ 1984-1985 major I ntchnge 0.9 to 6.1 1600
46 walls

Jasper Asphalt lithonia Mater lal

19 Plant GA 1985 handling 6.1 32
Operation platform

Gaspe 0 Gaspe 0 coastal

20 Pen i nsu I a Peninsula 1985 protection 5.3 99
Seawa II Quebec

Stump Sierra Natl. retal ned

21 Spr logs Forest 1985 fill 5.5 24
Road

Rush Angeles retained

22 Canyon NaTI. Forest 1985 fill 6.1 15

King County temporary

23 SR516 WA 1985 wall 2.4 30

King County temporary

24 SR522 Wa 1985 walls (5) 2.1 to 3.4 1372

Fish Creek Umpqua earth

25 Road Nat I. Forest 1985 re I nforcement
OR

Cascasde Dam Cascade crane operating

26 Hydroelectric MI 1986 platform 3.0
Facility

Interstate Glenwood retaining wall

27 1-70 Canyon 1986 for 4.6 61
CO roadway

St. Pau I retaining wall

28 TH 35E MN 1986 for 1.5 to 4.3 300+
roadway

Coa I bank Coos County temporary 1.8 to 3.0 30

29 Slough OR 1986 walls (2) 1.8 to 5.5 33
Br i dge

Coqui lie Coos County tempor ar y wa I I

30 River Bridge OR 1986 for detour 3.0 9.1
Detour

Lerroyne retained fill

31 Lemoyne PA 1986 for park i ng 4.0 82
lot

Pacl f Ic County temporary

32 SR 4 WA 1986 walls (2) 3.0 120
 243

Wrapped I: 4 Sandy cl ay
slope Woven Mirafi 600x Polypropylene si It backfi II
no facing

Wrapped Nonwoven
gun Ite need I epunched Supac 12NP Pol ypropy lane

Wrapped 1.5: 1 Tansar Pol yethy I ene S II pout

to I: I 5 lope Geogr I d SR2 HOPE cou ••d by 19
canpacted straw intense storms
. - - -------------------------------------------------------
fu II height
_. - --------------------------
first fu II he i ght
~

pre-cast concre geogr Id Tensar SR2 polyethylene concrete t i l t 20,21,22

pane I 5 HOPE up tace In U.S.

cruc i form use of geogr f ds

pre-c8st concra geogrld Tenser SR2 polyethylene wIlner wall 20,23
panes I 5 HDPE panels

concrete wave 1st N.. Amerl sea

deflector
panels
gaogrid Tensar SR2 polyeThylene
HDPE so,
wall w/polymer
based I re I ntoree
24

slope 1: 1 Tenser 551

CElli googr Id SS2 pol ypropy lane
Tensar on slope

slope 1.5:1 googr Id Tensar SS2

wrapped uv stab iii zed Permaa-tax granu I ar

no facing woven 5 lit 2300 pol ypropy I ene bockf III
film

wrapped uv stabi Ilzed Permea-tex granu I ar

no facing woyen s l i t 2300 pol ypropy lene backf III
film

googr Id Tenser

wrapped wood construee 4

facing attached googr I d S I gnode TNX250 po I vester days wI i nexper 25
after construct crew

non woven based on resu I ts

wrapped gun i te heat bonded Typar 3601 polypropylene of glenwood
canyon test wa II 5

uv stabilized Nicolon
treated wood woven Goo I on 500 polypropylene

"non woven 1 foot layer

need I epunched Supac aNP polypropylene spacing

I to 2 foot
layer spacing

wrapped· gun i te woven MI rl!lf i 1200HP pol ypropy lene

granu lar
bockflll
244

TABLE 1. GENERAL INFORMATION (cont'd)

CASE PROJECT LOCAT ION YEAR STRUCTURE HEIGHT LENGTH

No. NAME CONSTRUCTED (ml (m)

Snohom i sh walls temp (8)

33 SR 2 &, King Counties 1986 permanent (2) 1.5 to 6.7 490
WA

36th Avenue Anchorage Started reta i ned til I

34 Regrade A I aska 1986 for bike path 1.2 to 1.8 130

Oevon Devon Started

35 Geogrid Test AI berta 1986 test 5 lopes 11.9 72
Fi II

I 210 over Lake Char I es under retaining

36 Southern LA construction wal J 4.6 253
Pacific RR

I 20 over Talullah under re i n forced

37 Chicago mi II LA construction stope 3.4 to 6.1 1400
box cu I vert

NATO Advanced Royal Mi I itary testing in

38 research College Kingsto progress T9St walls (2) 2.7 2.4
workshop Ontar 10

North Ha I awe Oahu to be retaining walls

39 Va II ey access HA constructed for approach 3.0 to 7.3 854
road roadway

FHWA A I goqu j n under retaining

40 test wa II 5 IL construct i on walls 6.1 11

3 .t 6.1 11

6.1 11

6.1 11

2.t7.6 15

2 at 7.6 15

Phi ladelphia
41 Conra i I PA 1987 rai I line 6.7 90

Mary land Chr Istl na

42 Avenue DL 1986 park i n9 lot 1.2-3.7 110
 245

FACING RE I NFORCEt~ENT COMMENTS

TYPE BRAND POLY~R

wrapped uv stab iii zed granu I ar

permanent woven sl it Permea-tex 2300 pol ypr:.opy I ene backf i II
wallsgunite film

wrapped wood Mlraf I for support

f ac i n9 attached ge09r i d Paragr I d pol yathy I ene of bike
atter construct 100/255 path

Mlraf I attempt to
1: 1 slope ge09" I d Paragr I d 505/ po I vester determ i ne so i I 26
50S & 5T tabr Ie

S I gnode Interaction
geogrld TNX 5001 & pol vester mechan I sm for
TNX 250 cohesive soils

Tensar po I yethy lene

geogrld SRI & SR2 HOPE
551 pol ypropy lene

Bay Mills modified

9 losgr I d Glasgrld asphaltiC glas !o

wrapped po I yethy I ene exper I mental

gun i te geogrJd Tenser SR2 HOPE project

slope 4: 1 exper Imentel

no tael"9 googr Id Tensar SS2 pol ypropy lene project

I ncremente I pol yethy I ene exper imental

propped vert i ca googr Id Tensar SR2 HOPE project 27
timber panels

wrapped 1: 5 MOst extens I ve

slope reinforce non woven po I ypropy I ene use of tabr I c
gunite ,.e' nfore i"9
articulatIng rei nforced .v. I uate des I gn
pi eeast steel strips earth steel INfhology 28
concrete pane I 5

bar mats VSL steel

geogrld Tenser SR2

non woven
geotextlle Qullne polyester

IH: IV woven Amoco 2006

slope geotextlle Signode polypropylene
geogrld

IH:2V
slope

wrap around wI
precast coneret geogrld Tenser SR2 HOPE 36
bk fal se face

gab Ion geogrld Tensar SR2 HOPE 36

246

TABLE 1. GENERAL INFORMATION (cont'd)

CASE PROJECT LOCATION YEAR STRUCTURE HEIGHT LENGTH

No. NAME CONSTRUCTED eml (m)

Rottvdam stream
43 Rotterd~ NY 1986-87 diversion t!. 5.5 )80
parking lot

Weltertord Houston mar i na

44 Harbor TX 1986 bu I khead 2.4 3350

Mob! Ie driveway &

45 AL 1986 parking lot 1.8-3.7 74

Tyler land
46 Res i dence Inn TX 1985 deve lopment 6.1 104

Westbank prOTotype
47A Elliott Plaza British 1985 retaining 2.4 33
Columbia wall

478 2.4 33

Kelowna retaining
48 Smith British 1985 wall 2.4 3D
Col umble

Naramatta Pentkton
49A Road Br i t i sh 1985 1.8 29
Columbia

498 2.4 32

Gorman Westbank retaining

50 Kiln Br itlsh 1986 wall 1.8-4.9 44
Columbia

Gorman Westbank
51 Shop British 1986 3.7-4.9 32
Columbia

Borges Vernon
52 Rehabll itatlon British 1986 3.1-4.9 15
Col umbla

CNR Vernon
53 Overpass British 1986 2.4 162
Col umbl a

Kelonna
54 Torhlelm British 1987 1.2-6.1 99
Col umbia
 247

FACING RE I NFORCEMENT COMMENTS REFERENCES

TYPE BRAND POL YMER

full height
concrete geogr j d Tenser SR2 HOPE 36

full height
concrete geogr i d Tenser SR! HOPE 36

masonry
blocks geogr 1d Tensar SR2 HOPE 36

full height
concrete geogrld Tenser SR2 HOPE 36

po I yethy I ene
waf t I eerete geogr i d Tensar SR2 HOPE

polyethylene reinforce rear

watt I ecrete geogrid Tensar SR2 HOPE wall gf dry kiln
to support we I I

cast in rehab I I itetlon

place of f8111"9 cast
concrete In p lace con. wall

waf f I eerete

Tenser SRI
SR2
248

TABLE 2. DESIGN INFORMATION

CASE HE I GHT RE I tf'ORCEMENT RE I tf'ORCEMENT SUlOiNlGE FOUNDATION

SP N: I NG LENGTH LOAD
No. (m) (m) ho) SOIL STRATI FICAT1 ON WATER c 0
TYPE TABLE kPo

I mater of fill
3.0 0.23 - 0.30 3.0 duel tandem rock
axel loads

930 kN weath ..... d

2 0.9 - 6.1 0.23 - 0.30 4.0 on 4 axels rock

design 3.7 0.6 vertical

3 foll.d3.0 1.2horlz. 3.0 lean clay

311 kN
4 6.7 0.23 4.0 dual tandem rock
whee' load

stiff to hard high water

4.9 0.15 - 0.23 3.7 silty cloy wi un! form table
sand & grave I

4.0 at base 311 kN old

8.5 0.3 11.9 at top dual tandem I ends I ide 38
wheel loed

highly compress uniform,

9 4.6 0.23 - 0.40 4.0 compress 1b I e th' ck depos I ts
5 II ts and c I'ays

clayey silt high Wider

10 4.6 & 5.5 1.2 3.7 - 5.5 30 kN/m till table 6.9 31

silty high water

II 6.7 0.61 0.9 Beaumont clay table 20 15

dolomitic
12 8.2 0.2 - 0.4 3.4 limestone
bedrock

soft gray shale-fractured

13 9.1 0.3 - 0.9 4.9 sha I e and weathered

14 11.9 0.4 - 0.6

13.6 kPa & ,

16 3.0 0.20 2.4 160 kN dual si Ity sand un I torlt 30
tandem wheel

17 14.0 0.61 6.1

col laps ib Ie
18 0.9 - 6.1 0.3 - 0.9 3.7 12.0 kP. soi Is

soft clay wI
19 6.1 0.3 - 0.6 3.7 12.9 kPa crushed stone 24
on top

dense sand
20 5.3 0.3 3.5 21.0 kPa and gravel at surface 0 30

-~---------------------------------------------------- ------------------------------------------------------
 249

REINFORCED Fill BACKF III

d w SOil CO 0° d w SOil CO 0" d \Ii

kN/m3 1%) TYPE kPa kN/m 3 1%) TYPE kPa kN/m3 1%)

si Ity sand, si ity sand,

gravel 5 i ze 38 18.5 gravel size 38 18.5
rock fr agment-s rock fragments

un i form crushed un I form

basa 1t 31 19.6 crushed basa I t 31 19.6

clean uniform
concrete 36 15.3 4.5 lean clay
sand

75mm minus 75rrvn minus

crushed rock 35 17.3 crushed rock 35 17.3

open graded open graded

crushed stone 33 crushed stone 33

wood ch j ps 32 5.8 wood ch j ps 32 5.8

rounded,well
graded clean 35 20.4
sand & grave I

granu lar 35 19.9

si Ity si ltv
Beaumont clay 20 15 Beaumont clay 20 15

25rrvn crushed si lty clay 144 10 22.0 16

limestone 40 )8.1
native soil 9.6 30

crushed basa I t 40 22.0

on-site si ltv
clay 24 20

25nrn crushed 25mm crushed

gravel 35 21.2 gravel 35 21.2

granu I ar 2.4 32

granu iar 34 19.6 granu I ar 34 19.6

granu lar 40 21.2 granu I ar 40 21.2

we II graded we II graded
sandy grave I 34 21.0 sandy gravel 30 2).0
250

TABLE 2. DESIGN INFORMATION (cont'd)

CASE HE I GHT RE II-FORCEMENT REII-FORCEMENT SUlCHARGE FOUNDATION

SPACING LENGTH LOAD
No. (m) (m) (m) SOIL STRATI FICA TI ON WATER c
TYPE TABLE kPa

1.8 3.0 landslide

21 5.5 --------------------------- H-20 loadl ng SM-GM deposit 34
1.8 0.9

22 6.1 0.76 4.0 H-20 loadl ng SM-SC un t form 24 20

2.4 at base 840 kN crane, medium dense

26 3.0 0.3 4.6 at top 180 kN lifting fine sand wi 30
capacity some clay

un i form bedrock 33 -
28 1.5 - 4.3 0.3 - 0.6 clay at 3 to 4.6 m 57 20

15.9 kP. & shale rock

31 4.0 0.15 160 N dual fragnents & un I form 20
tandem axle 5i It

34 1.2 - 1.8 0.61 2.9 straTi fled

primary 2.0 pr IlIary 13 soft sl tty at minus

35 secondary 1 secndry 3-5 clay strati f jed 6.1 m 20 27
terTiary 0.3 tertiary 1.5

at minus
36 0.23 - 0.61 3.8 clay strati tied 0.3 m 50

at minus
37 3.4 - 6.1 0.61 varIes clay strati tied 2.1 m 100 0

0.76 m of sand sand

38 2.7 0.76 3.0 & lof Istable blinding 0.24 m thick 43
air bags layer

weathered
39 3.0 - 7.3 0.15 - 0.30 2.3 - 3.4 rock

gravely at minus
40 6.1 0.38 & 0.76 4.3 3m sand sand un I form 0.5 m 0 39.0

7.6 0.76 4.3

gravely
sand
.
un I form
at minus
0.5 m 0 39.5

41 6.2 rai troad

42 1.2-3.7 0.91 parking JOT

GW 1st 1m
43 5.5 0.3-0.6-1.2 4.4 12 kPa GW clay & silt 45

44 2.4 3.2 4.8 kPa CH pil log below 6.2 16

46 6.1 0.2-1.2 4.6 o kPa SIH~L 30

 251

RE I NFORCED FILL BACKFILL

d w SOIL CO 0° d w SOIL CO 0° d w
kN/m 3 (%) TYPE kP. kN/ml (%) TYPE kP. kN/m 3 (%)

18.8 SM-GM 34 18.8

18.1 SM-SC 24 2D 18.1

fine to coarse medium dense

sand &. gravel 37 19.6 fine sand '11/ 30
some cl ay

30- 30-
granu lar 35 18.1 granular 35 18.1

sand &- gravel, shale rock

2).2 -- 10% tines 3521.2-- fragments & 20 21.2
si It

75mm minus
sand &- gravel

inorganic clay 21.5 inorganic clay 21.5

35 of 10 to medium 13 22 15.7 Of.(;: of low to med 13 22 15.7 OMC
plasticity plasticity

inorganic clay
15.7 25 sand o 30 -- of low to med
plasticity

5 0
13.3 37 clay ---------- 13.3 --
o 12

washed san d '11/ washed sand w/

17.0 some gravel ,41 16.03 some gravel 41 16.0 3

weathered
granu lar rock

gravel o 42 18.0 gravely

20.2 sand o 39.0 20.2 2 sand 39 20.2 2
sl It 13 35.517.0 19

gravely gravel y
20.2 samd o 39.0 20.2 2 sand o 39 20.2 2
sl It 13 35.5 17.0 19 sl It 13 39.5 17.0 19

Sl~ 35 18.1 -- SM 30 18.1

CH 6.2 16 19.2 -- CH 6.2 16 19.2 --

18.8 -- SM-ML 30 18.8 SM-ML 30 lB.8 --

252

TABLE 2. DESIGN INFORMATION (cont'd)

CASE HE I GHT RE I NFORCEMENT RE I NFORCE~NT SLRCHARGE FOUNDATION

SPAC ING LENGTH LOAD
No. (m) (m) (m) SOIL STRATIFICATION WATER c
TYPE TABLE kPa

47A 2.4 0.6 & 1.2 1.4 GP variable o 33.5

2.4mwall
47B 2.4 0.6 & 1.2 2.8 above GP variable o 33.5

fine river
48 2.4 0.6 & 1.2 1.8 SP sand -1m

49A 1.8 0.6 1.1 GP

496 2.4 0.6 & 1.2 1.3 GP

4.8kPa
50 1.8-4.9 0.6 & 1.2 2.9 of backf III SP 35

B.C.F .S.
51 3.7-4.9 0.6 & 1.2 2.9 L-45 GP 35

52 3.0-4.9 0.6 & 1.2 2.9 4.8kPa

12kPa backf III

53 2.4 0.6 3.0 33.7 degree SP varleble 35
slope

54 1.2-6.1 0.6 3.2 2.4kPa SP varved 3D

 253

RE INFORCED FILL BACKFILL

d w SOIL CO 0° d w SOIL CO 0" d w

kN/m! 1%) TYPE kP. kN/m! 1%) TYPE kP. kN/m! 1%)

20.4 -- GP o 33.5 20.4 -- GP a 33.5 20.4 --

20.4 -- GP o 33.5 20.4 -- GP o 33.5 20.4 --

SP o 33.5 19.6 -- SP o 33.5 19.6 --

GP 32 17.3 -- GP 32 17.3 --

GP 32 17.3 -- GP 32 17.3 --

20.4 -- SP 35 20.4 -- SP 35 20.4 --

20.4 -- GP 35 20.4 -- GP 35 20.4 --

19.6 -- GP 35 19.6 -- GP 35 19.6 --

18.8 -- SP 30 18.8 -- SP 30 18.8 --

254

TABLE 3. REINFORCEMENT PROPERTIES AND DESIGN METHODOLOGY

RE IKFORCEMENT PROPERTIES DESIGN PROPERTIES DESIGN /!ETHOLOGY

CASE METHOD OF TENSilE MAXIMUM STRENGTH Ie REOUCTJON FACTORS EXTERNAL INTERNAL STABILITY
No. STRENGTH STREKGlll ELONGAT 100 ---------------------------- 5TABIL I TY
EVALUATION kN/", IS) TlkH/JIIJ (S) DLRAaIU TY CREEP MODEL VERT. STRESS H.STRESS OISTR PULLOUT

flb..-t •• 2'- strip 51 [ding tieback

"
'00 te$t 12.) o ....r-turnln g wedge overburden 2/3 0

".
Flb.-tell 25_ strip sliding tieback
'00 test 1 1.4 O\I.,.turnlng ",edge overburden Ko 2/3 0

Flb..-tex
600 17.1

Bldl.
C-28 10.7 60

------;j';;;--------:;----------------------------------------------------------------:--------------
C-Ja 18.9

heavy duty webbing c apston t ieback

nylon grip tST "RIOd 190 14.5 .,edge Ko and Ke
strips greb/501llll wide

Flbertex 2'1WII strip test 5.2 '.2 slidIng

ZOO l00rm strip t5t
grab test
8.4
1605 N)
8.'
(605 Nl
overTurning

Bldllll Tieback
C-34 greb test (1000 Nl 1J.1 wedge 2/3 0

Supac greb test 1670 Nl slIding TIeback

5W ------------------------------ 12. 3
strip test 23
overTurning
landslide
wedge overb urd en '0 2/3 0

'.8
,..
Flbertex 2'00 wid e wIdth tST 140 tIeback
9 ------- w- 2oonn wedge
Fl bertex 400 1- 100II1I'I 10.0 14'
Supac 41f' 12.6
" '.1
...
t-l01 ----------------------------------------
Supac 6If' 24.3 60 9.8

'.7 80

Typar }401
16.6

7.7
"
60
10.9

3.2

" 4.2
Typer :5601 12.4

" 12 \6.1 global 0.8 ton 0

menutecturer sllde/oyerTurn
12 Supac 121f' 25II1II grab
t.ST
(1470 Nl 12.8 short/long Term
slope STo!Iblllty
overburden
'0 pu II out factor

OC sing le rib
1:5 Tensar SR2 W-', L-3 rIbs
t-50".,/1II1n
79 12 15.8
40'
manufacTurer sliding
16 Supac 12'" 25l1li grab
test
11470 IU 12.8 bearing
cap&c:lty
overburden
'0 pu I lout factor

'.7
'"
17 T.n5&111" SR2 glQbeli pullout test

Index test sliding bearing tieback

18 Tenser SR2 'If-I',L-' rIbs 70 16-2' 10 overturning wedge trapezoida l retio 0.9 ten 0
1'*21/lIIln global stabiliTy

Index test sliding bearing Tiebec::k

19 Tenser SR2 W;oI5,L-' ribs 70 16-25 29 10 overturning wedge trapezoidal ratio 0.9 tan 0
t-21/_1 n global stllbll ltv
 255

TABLE 3. REINFORCEMENT PROPERTIES AND DESIGN METHODOLOGY (cont'd)

R£ I /,FOACEMENT PROPERT I E5 DESIGN PROPERTIES DESIGN II£THOLCX;Y

CASE toETHOO OF TEN S ILE ,...... XIMUM STRENGTH @e REDUCTION FACTORS EXTERNAl INTERNAL STA6IL ITY
STRENGTH STRENGTH ELONGATION -------------------"---------- STAB ILITY
EV,lJ.UATION kN/m (~l TikN/m) IS) O~A6ILITY CREEP Po4COEL VERT. STRESS H.STRESS OISTR PULLOUT

10000 hr. creep s l idil'l9 bei!lring tie-bK.k

20 Ten5~r SRZ tst edap. t o 29 21.5 OVe!'" tu r n I 1'19 wed ge 0 .9 tan 0
70 yr. design g lobi!ll stab I I ity

TenS8r 551 manufacturers incorp. sl idlng bearing pseudo

21 i!IIIId SS2 Ii ter8ture 12.8 I design oYef"turning tie-back /oteyerhot
strength ... edge

manufacturers incorp. sliding bellring pdeudo

22 TenS1I1' SS2 II teraTure 12.8 I design over turning tie-back Myarhot
strollgth wedge

Signode wide width 5 1 tdlng be!lrlng tie-b!lck Myerhof 0

26 TNX2~ ASTM 0-4595 48 '.3 "3 1/3 overturning
deep seat slope
wedge and
Trapezoida l
K-5tres5 ratio pullout test

wide width
,., t ie-back
21 Typar 3601 w" 200lm,Lz10Omm
t ,. 10:{
12.4
" wedge

Nlcolon wide width 29 10 sliding bearing

28 Geolon 500 -.---------------------------------- 26.2 10 capoaci ty
grab test 35

IoIlrat1 Mi'lnufacturer sliding bearing tie-back

31 1200 liP 1"",1" 19 11.5 cepaclty wedge Ko pu I lout factor
Jaw test

10000 hr., 5f Idlng-I.S tieback

43 Ten sar SR2 1000000 hr.
e"'trapolated
19 ro '.0 1.0 beitl"lng-2.0
overturnlng-2 .0
wedge Mey.,.hof K. test

slldlng-I.5 tieback
44 Tenser SR2
" 10 29 '.0 1.0 bearing-I.,
ov erturn lng-I.,
wedge Trapezoidal test

slldlng-1.5 tieback
'.0
46 Tenser SR2
" ro 29 1.0 be arlng-2.0
overturn! ng-2.0
wedge Trllgezoldal K. test

41 Mfg. bi!lsed on Incorp sliding beerlng tlebeck

thru Tensar SR2 1~ or less 16.1 design cap. overturn wedge loIeyerllo f
53 creep In 120 yrs strength sl ip tal lure

54
Tensat' SRI
SR2
IoItg. based on
IS or less
8.' Incorp 5 1 tdlng beat'lng tieback
design cop. overturn wedge Meyer~f .
creep In 120 yrs 16.1 stren gth stlp failure
256

TABLE 4. INSTRUMENTATION

LATERAL MOVEMENT OVERALL SeTTLEMENT

CASE -----------------------------------------------------------------------... ------------------------------
No. STRAIN IN AND INCREMENTAL SURFICIAL FOUNOAT I ON
IN FACE BEHIND SETTLEMENT SETTLEMENT SETTLEMENT

.
_ _ _ _ _~_ _ _ _ C _ _ _ ~_ _ _ _~______I~:_~~_ _ _ _ J::~_ __
RE I NFORCEMENT STRUCTLRE oF WALL OF WALL (differential J

26 hor i zonta I 26 horizontal

transit lncllnaaeter lncl inometers level level
tubes w/sllp rings

6nvn to 52 rrm o to 52111f1 15 to 30rrm

indirectly
calculated observation

bottom 57 Iml
top 152 "'"

slope indicator
slip tube & settlement
devices

o to 6.3rrrn 6.3mn to 33.5rrrn

survey posts 30 her 1zoo ta J 5 vertical lnclln

and 18ser I nc I I naneter- 5 manometers observational -ometers sondex
targets extensometer 5 manometers

O.3m over 90m large amount

length after of consolldetlon
first 3 roonths O.6m at 1'/. end wall

survey survey
reference extensoneters I nc I 1nometers reference I nc II nometers
points points
17 ------------------------------------------------------------ ---------------------------

reslstence
strain gauges observational
& Induct.colls
16 ----------------------------------------------------------------------------------------------------
line of mex
strain 18 to 65 II1II
2' frc. v.-tlcal
res I sT.nee
strain gauges obs.. v .... lonsl
& Induct.colls
19 --------------------------------------------------------------------------------------------------
data "i'thin
inconclusive anticipated
range
 257

TOTAL STRESSES
INTERNAL LOAD QiANGES Of LOAD PORE PRESSURE TEMPERATURE COI+IENTS
BASE OF FACE OF STRESSES DISTRIBUTION DISTRIBUTION RESPONSE (I f i nstrOOlentction
WALL WALL (soil) IN REINfORCEMEN WITH TIlE BELOW WALL

!:_ _ _ _ _ _ _ _ _ _
~ffected)

___~_____~______~______~______~_____[E ___________

6 wES pressure 6 WES eXCf;lSS i ve

ce I Is-center pressure deformat i on of
line lft.trom face cells the reinforcing
------------------------.--------.-------------------------------------------------------------- str ips beyond
measured pressure that necessary
I ess than pred 1cted to create
except I!It fa i I ure Rank I ne steta

load cells pi esometers

loads larger excess pore

near back press. did not
of wall bui Id at loading

resu Its of the

i nstrumentat ion
--------------------------------------------------------------------------------------------------------- program were
not avai lable

18 SINCO
load cells load cells load cells strain gauges strain gauges res i stance monitoring
thermom probes program
sti I I in
approached approached R2Ink i ne wedge Rank i ne wedge progress
pred i ct press. pred i ct press. def i ned def i ned
at toe of wall at toe of wall

load cells load cell s load cells strain gauges strain gauges monitoring
program
sti II in
approached approached data data progress
pred i ct press. predict press. inconclusive inconclusive
at toe of wall at toe of wall
258

TABLE 4. INSTRUMENTATION (cont'd)

LATERAL MOVEMENT OVERALL SETTLEMENT

CASE ----------------------------------------------------__ ----_____________________________________ _
No. STRAIN IN AND INCREMENTAL SLRfICIAL FOUNOATI ON
IN FACE BEHIND SETTLEMENT SETTLEMENT SETTLEMENT
RE I ~ORCEMENT STRUCTURE OF WALL CF WALL (differential)
------------------- -----,----------------------------
~r-~~rc.4
------------..-
------------------------------------------------------
stre I n gauges mutl position
... ---------~---

glued to the ob~ervatlonaJ mach. ex.tenscrn observational

re i ntorcement & observation
20 ---------------------------------------------------------------------------------------------------
0.75% 47 .... 37 .... 17rrm to 3611111

m I crones urement wire

& bison Induct observational extensorneters observational
stral n gauges
26 --------------------------------------------------------------------------------------------------
1% upper level
0.85% lower level
of wall

elect resistanc horizontal horlz. extensDm

vert. extensom vert. extensom vert. magnet I c
& b I son stra I n extensometer and verticaland horiz. and horiz. extensom &
~ gauges I ne 11 nocneter i ne I Jnaneter i ne I i naneter I ne I i nameter
35 ----------------------------------------------------------------------------------------------------

strain gauges Inclinometers

36 -----------------------------------------------------------------------------------------------------

strain gauges i nc I i nometer s

37 -----------------------------------------------------------------------------------------------------

observat lona I observati ona I

38 ----------------------------------------------------------------------------------------------------
 259

TOTAL STRESSES
INTERNAL LOAD CHANGES OF LOAD PORE PRESSURE TEMPERATURE COMMENTS
BASE Of FACE OF STRESSES DISTRIBUTION DISTRIBUTION RESPONSE (It instrumentation
WALL WALL <5011) IN REINFORCEMEN WITH TIME BELOW WALL affected)

peter stress six peter

cells stain gauges strain gauges pneLmatic thermoooupies
pi ezaneter 5

rankine tal lure

erratic results surface no conclusion
def I ned

load cell at
tie rod location strain gauges strain gauges wall p~rformed
sat i s factor y
----------------------------------------------------------------------------------------------------------- d ur I n9 cr ane
loading
no conclusion no conclusion

20 S I NCO frost heave

stre i n gauges stra 1n gauges pneumati c gauges with construction
pi ezcmeters thermocoup les not
campi ete

plezerneters

@ panel
load celts geogr 1d stra I n gauges stra I n gauges thermocoup les
locations
260

TABLE 5 TENSAR GEOGRID REINFORCED SOIL RETAINING WALLS

NORTH AMERICAN SUMMARY

FACING: WRAP- GABION

AROUND

APPROXIMATE NUMBER OF
WALLS CONSTRUCTED
(+ or -30%): 40 10

STRUCTURES: Temporary Const. Land Development

Land Development Erosion.Control
Blast Protection Walls over high
Railroad support settlement areas
Walls over high Landscaping
settlement areas
Landslide repair

TYPICAL HEIGHT RANGE: 8' to 30' 8' to 30'

REINFORCEMENT - GEOGRID

BRAND: a. Tensar SRI, Tensar SRI, SR2,

SR2, or SR3 or SR3
b. Tensar SS2

POLYMER: a. HDPE HDPE

b. Polypropylene

TYPICAL REINFORCEMENT 8" to 24" 18" or 36"

SPACINGS:

FOUNDATION SOILS: Any soil type Any soil type

NOTES:

Embedment Lengths: L=0.6 to 1.0 x Height; L=f (soil strength, sur-

charge loads, etc.); Typically L=0.7 x Height

Reinforced Soils: Any Soil Type, Excluding Organics

Retained Backfill Soils: Any Soil Type, Excluding Organics

Design Tensile Load: Approx. 40% x Ultimate Strength for SR products

and approx. 25% x Ultimate Strength for SS pro-
ducts prior to factoring

Reduction Factors: For Durability 1.0, Creep 1.0 as it is accounted for

in determining Design Tensile Load, Site Damage
typically 1.0, Range from 1.0 to 1.35 Used
 261

TIMBER MASONRY FULL-HT. ARTICULATED

PRECAST CONCRETE OR
CONCRETE MASONRY UNITS

10 15 30 16

Landscaping Land Development Highway walls Highway walls

Land Development Highways Land Development Land DevelQP-
Bridge wing walls ment
Harbor walls Walls over
high set.tle-
ment
Seawalls

5' to 18' 4' to 24' 3 ' to 20' 12' to 24'

a. Tensar SR1, Tensar SR1, SR2, Tensar SR2 and Tensar SR2
SR2, or SR3 and SR3 SR3 and SR3
b. Tensar SS2

a. HDPE HDPE HDPE HDPE

b. Polypropylene

7", 14", or 21" 8", 16", 24" , 8" to 4'0" 12" to 24"
or 32"

Any soil type excluding organics.

External Stability: Typical Factors of Safety - Sliding: 1.5

Bearing: 1.5 to 2.0
Overturning: 2.0
Global: 1.3 to 1.5

Internal Stability: Mode - Tie-back wedge

Vertical Stress: Trapazoidal distribution approximately through 1985,

Meyerhof type distribution since then

Horizontai Stress Distribution: Ka

262

partially instrumented, of which four were research walls and

nine were production walls. Major research projects that are
currently underway (Projects No. 35, 38 and 40), but have not
been completed, were also included in the summaries, as they
will undoubtedly complement the currently availabile
information.
It is believed that the projects surveyed are only a portion
of the total number of walls and slopes reinforced with
polymeric materials in North America. For example the
attached Table 5 provided by the Tensar Corporation presents
a summary of structures where their products have been used
and includes an estimation of 121 projects. Likely there are
numerous other unreported reinforced soil projects where
other geosynthetics have been used. Our best estimate of
polymerical reinforced soil walls and slopes that have been
constructed in North America is on the order of 200.

The' following section provides an overview of projects that

were either unique or where significant technical information
was available.
2.1 Overview of Technically Significant and Unique Projects

The Illinois River Road Wall project, constructed in 1974 in

the Siskiyou National Forest, Oregon, was the first reported
full-size geosynthetic retained soil wall built in North
America.' The United States Forest Service [1] designed and
built the wall to restore the width of Illinois River Road
which had been reduced by erosion. The 2.7 m high and 10 m
long wall was designed using geotextile reinforcement. A
conversation with the designers indicates that the wall is
still in service and has performed satisfactorily since
construction in 1974.

Between 1974 and 1976, the United States Army Engineers

waterways Experiment Station in Vicksburg, Mississippi
investigated the effect of horizontal reinforcement on the
stability of earth masses [5, 6]. Two instrumented retaining
walls were constructed. The walls were to b~ 3.7 m high, 4.9
m long and 3.0 m deep. One wall was reinforced with neoprene
coated nylon fabric strips and the other with galvanized
steel strips. The fabric strip reinforced wall failed when
the wall height reached 3.0 m. The exact cause of the
failure was unknown, but is believed to be the result of
significant elongations in the reinforcement beyond that
which was needed to create active failure in the soil. There
was also some indication that the elongation could have been
concentrated at the face connections. Lateral pressures were
measured at the face of the wall. The measurements indicated
lateral pressures slightly less than those predicted by
Rankine theory prior to failure and equal to or greater than
the predicted lateral pressures at failure.
 263

In 1981, the California Department of Transportation built a

tire-anchored timber wall on state Route 203 in Mammoth Lakes
(9). The tire-anchored timber wall incorporated two waste
materials - used automobile tire side walls and used railroad
ties. Two walls varying in height from 1 to 3.7 m and
totaling more than 230 m in length were constructed. The
walls performed far better than expected, including
withstanding an earthquake of 5.8 Richter magnitude in 1981
with no visible evidence of damage. Several other walls were
constructed by Cal trans using the same system as the Mammoth
Lakes wall. The walls were constructed in the following
locations in California: Route 101 in Marin County (1983),
Route 1 in Santa Cruz County (1983) and Big Basin Park in
Santa Cruz County (1984). No other information was available
concerning the projects.
The first geogrid reinforced wall constructed in the united
States served to stabilize a landslide near the entrance to
Devil's Punch Bowl State Park in Oregon in 1983. The
project, a FHWA Experimental Features Project, was to assess
the construction problems of near-vertical walls with
geogrids. The geogrid wall was chosen over the other
alternatives for two reasons [16, 17):
1. It had the lowest estimated cost $260/m 2 , and
2. The open face structure of the geogrid allowed
establishment of vegetation, which provided a
natural appearance compatible with the surrounding
state park.
The geogrid wall was less expensive than the geotextile wall,
since the geotextile wall required a protective facing.

In 1982, the Colorado Division of Highways authorized the

construction of a 4.6 m high, 91 m long geotextile reinforced
soil wall test section in Glenwood Canyon, Colorado [9, 10).
The 91 m long wall was divided into ten, 9.1 m sections, each
of which incorporated the use of a different style or weight
of non-woven geotextile. The purpose of the test wall was to
determine the validity of the design methodology in
evaluating the properties of the reinforcement, including
creep. All sections of the wall were fully instrumented as
indicated in Table 4. Two sections of the test wall were
intentionally underdesigned. These two sections were then
surcharged with 5.2 m of fill in an attempt to produce
failure; however, under the full height of the surcharge,
insignificant internal and external deformations were
observed. As of 1975, creep related movements have not been
obse rved in any of the wall sections, even though the soft
sil t and clay foundation soils have been compressed
approximately 0.6 m at one end of the wall [11]. Fabric
samples have been taken on a yearly basis and tested for
strength· loss by Professor Bell of Oregon State university.
264

The results of the testing program were not available at the

time of this paper. In 1986, the Colorado DOH constructed a
geotextile reinforced wall in conjunction with the 1-70
project through Glenwood Canyon. The wall, designed in
accordance with the findings of the 1982 test wall, has
performed satisfactory to date.

A series of high densi ty polyethylene (HDPE) geogrid

structures using full height concrete tilt up facing panels
were constructed between 1984 and 1985 in Tuscon, Arizona
[20. 21, 22]. A total of 46 walls ranging in height from 1 m
to 6.1 m and totaling over 1600 m in length were constructed
for a series of grade separations for a major thoroughfare.
Two wall sections were heavily instrumented and only limited
results have been reported to date. A preliminary assessment
of lateral and vertical pressures indicat~d that the. lateral
pressures at the face of the wall were less than those
predicted by either the Rankine, Meyerhof or trapezoidal
pressure distribution theories. However, vertical pressures
at a distance of approximately 0.3 m behind the wall were
found to exceed those predicted by either the trapezoidal,
Meyerhof or overburden pressure theories. Local strain
measurements on the reinforcement indicated a wedge type
distribution of peak strains over the height of the wall that
agreed with a Rankine type failure surface.
The first reported polymer-based reinforced soil sea wall in
North America was constructed along the coast of the Gaspe'
Peninsula in the Canadian province of Quebec in 1985 [24].
The 5.2 m high and 100 m long sea wall protects Highway 132
and the adjacent land from storm waves wtiich can reach
heights of 3 m. The sea wall was constructed using sectional
articulating type precast concrete facing panels and high
density polyethylene geogrids. An extensive monitoring
program was established to aid in evaluating the performance
of the sea wall. The sea wall was reportedly performing as
anticipated after 1.5 years of service.

To allow access of equipment for reconstruction of the

Cascade Dam hydro-electric facility in Cascade in Cascade,
Michigan, the owner, STS Consultants, Ltd. designed and
constructed a reinforced soil wall using polyester geogrids
to increase the width of the crest of the dam [25]. The soil
wall was required to support a 840kN crane with a 100kN
lifting capacity. The wall constructed in 1986 was fully
instrumented to monitor internal and external lateral
movement as well as strain levels in the geogrid elements.
The wall was monitored during construction and over the five
month period following construction up until and including
the operation of the high capci ty crane. Preliminary data
indicated no apparent strain over the five month period
fol~owing construction. The strain level during external
loading by the crane was found to be close to that predicted
by theory at the base of the wall, but exceeding that
 265

predicted by theory in the upper levels of the reinforcement.

Monitoring is being continued, so that long-term performance,
including stress relaxation after removal of the crane, can
be evaluated.
Several future projects of interest were also reviewed. More
than 39 kilometers of approach roadways will be constructed
in conjunction with the Interstate Route H-3 project on the
Hawaiian Island of Oahu [28]. The interstate will connect
the northeast coast of the island with the rest of the
island. The west access road alone, referred to as the North
Halawa valley access road, will requi re 6.3 kilometers of
retaining walls ranging in heights up to 8 m. The North
Halawa valley access road, when constructed, will be the most
extensive use of geotextile fabric retaining walls to date.
Environmental concerns have currently delayed construction~

Presently under construction is a 12 m high, 42 m wide and 72

m long geogrid reinforced test slope in Devon, Alberta [26].
The slope is constructed at 1 horizontal to 1 vertical with
native cohesive soils. Three test sections were established
to compare the performance of three different geogrid
materials. The test sections were desgined with the. minimum
number of geogrid layers and minimum length of reinforcement,
so that each reinforcing layer acts independently. The
primary objective of the test fill is to evaluate the
mechanisms whereby geogrid reinforcement can control lateral
strains in a cohesive soil mass.
Another series of test walls is being constructed under the
sponsorshiop of the FHWA in Algonquin, 11 [20]. The test
sections will consist of six, 6 m high by 11 m wide walls and
four, 7.6 m high by 15.2 m wide slopes. The walls will be
constructed using four types of reinforcement (metallic
strips, metal bar mats, a polymeric geogrid and a needle
punched non-woven geotextile). The soil will principally be
free draining granular materials wi th soil variation
evaluated for the bar mat reinforcement. A woven geotextile
and a geogrid will be used to construct 1 horiz.ontal to 2
vertical slopes with gravelly sand and 1 horizontal to 1
vertical slopes using silt. The structures will be fully
instrumented to monitor the stress distribution during
construction and subsequent external loading. The project is
to be constructed in the summer of 1987 with monitoring
completed by the summer of 1988. The project will conclude a
three year study to develop reinforced soil design
guidelines.

A test wall is also being constructed in conjunction with the

NATO workshop [38] for which this paper was prepared. The
resul ts of this test should be reported elsewhere in these
proceedihgs.
266

3. CURRENT TRENDS IN RETAINING STRUCTURE DESIGN

This next section describes current trends in the design of
geosynthetic retaining structures. The following factors
will be discussed:

1. Reinforcing Materials
2. Facing Systems
3. Backfill Materials and Interaction Mechanisms
4. Design Methodology and Factors of Safety
5. Project Costs
3.1 Reinforcing Materials
Of the projects surveyed, nine utilized woven geotextiles,
twelve utilized non-woven geotextiles and thirty-two ·utilized
geogrids. Other. polymeric systems that were used included
heavy duty nylon reinforcing strips and tire sidewalls with
anchor bars. From the introduction of fabric reinforced
retaining structures in North American in 1974 (Project No.
I), non-woven geotextiles were primarily used until the
introduction of geogrids in 1982 (Project Nos. 11 and 15).
Woven geotextiles were found to have had limited use as wall
reinforcement. Their primary application has been in
temporary wall construction.
The current domination of geogrids is probably due to several
factors. Firstly, the Tensar Corporation, the principal
geogrid manufacturer, has provided a strong technical
approach in its promotional efforts, an approach which had
previously been absent from the geotextile manufacturers.
Secondly, the apparent high strengths of geogrids may have
appeared attractive to designers, because they could reduce
the number of layers of required reinforcement.

The ability to use thicker lifts when geogrids are utilized

perhaps appeared attractive to the contractor from the stand
point of reducing construction time. However, as noted in
the Glenwood Canyon project [9, la, 16, 17) the forming
systems which had previously proven adequate ' for geotextiles
were not stiff enough to handle the large lift thicknesses.
Therefore, new forming systems were required. These new
forming systems were much more complicated and required
substantially more time to reset from one lift to the next
[16, 17). Thus, time saved because of the reduced number of
lifts was absorbed by the increased forming time.
Other factors influencing the choice of reinforcement include
the unit costs and the need for protection of the face
material. From the projects reviewed, geotextiles tend to
have a lower unit cost than geogrids as summarized in Section
3.5. Geogrids, apparently due to their thickness, have been
found to be less susceptible to UV deterioration than most
geotextiles [29].
 267

It is expected that newer higher strength geogrids and

geotextiles will permit the construction of even higher walls
than previously constructed using geotextiles. Of the
reported projects the tallest geogrid reinforced wall was 9.1
m, the tallest geotextile reinforced wall was 8.5 m and the
tallest geogrid reinforced slope was 14 m.

3.2 Facing Systems

All geosynthetic materials experience a loss in strength and
ductility when exposed to ultra-violet radiation. Therefore,
it is necessary to use either Uv-stabilized synthetics for
temporary applications or some other type of permanent facing
system for long-term applications. The facing systems used
in the projects evaluated included: gunite, emulsified
asphalt coatings, latex paint, timber, aluminum panels,
natural vegetation, and sectional articulating type and
full-height concrete panels. The most common system used for
geotextile reinforced walls was gunite. Gunite appears to be
a very satisfactory facing material. The gunite facing
system utilized for the Glenwood Canyon test wall remained
intact even after the wall experienced differential
settlements in excess of 2 ft. Guni te facing systems also
protect the geosynthetics from vandalism. The emulsified
asphalt coatings do not always adhere to the geotextile [8].
For example, the Camp Hill project utilized a woven slit film
geotextile which was to be coated with an emulsified asphalt;
however, when the asphalt failed to adhere to the fabric, the
fabric face was then painted with latex paint.
Several geogrid walls and slopes were constructed using
vegetation for cover to reduce uv exposure. As indicated in
project No. 18, full height concrete panels have also been
used. One innovative approach (Project No. 26) which can be
applied to both geotextile and geogrid walls is the post
construction attachment of facing panels.
3.3 Backfill Materials and Interaction Mechanisms

A majority of the 54 projects surveyed utilize' d granular

backfills. Granular backfill was used for all vertical wall
projects. Six of the slope projects utilized cohesive
backfills and two slope projects used wood chips.
Information was not available for 11 projects.
Although six of the projects surveyed utilized cohesive
backfill materials, unfortunately none of the projects
appeared to be instrumented and long-term performance data
was not available. Hopefully, the findings of the Devon Test
Fill project will lead to a better understanding of the
soil-reinforcement interaction for such structures.

In determining the pullout resistance for the projects

surveyed. several different approaches were utilized in
268

determining the soil-fabric friction angle. The approaches

included using two-thirds of the angle of internal friction
of the soil (f3), (0.8 to 0.9 tan(f3)), and the full
soil-reinforcement interaction angle obtained di rectly from
pullout tests.

3.4 Design Methodology and Factors of Safety

The tieback wedge method of analysis using a Rankine failure

surface was the only reported method used in analyzing the
wall structures. Intenral vertical stress distributions were
determined using either the trapezoidal, overburden or
Meyerhof stress distributions. A stress ratio determined
from an assumed bilinear failure surface after Jewell (1984)
was used in several geogrid slope cases. Both active and
at-rest earth pressures were used in the design of r 'etaining
structures with flexible facing materials.
Table No. 6 presents the factors of safety applied when
evaluating external and internal stability for the various
projects. The range and average of the factors of safety
used in the design of the various permanent walls and the
slopes are included in the table:

TABLE NO. 6: SUMMARY OF REPORTED FACTORS OF SAFETY

Stabili tJ': Factor Factor of Safety

1. Sliding: Range 1.2 to 2.0, Average 1. 54

2. Overturning: Range 1.2 to 2.8, Average 1. 66
3. Bearing Capacity: Range 1.5 to 2:0, Average 1. 89
4. Global Stability: Range 1.3 to 1. 5, Average 1. 40
5. Reinforcement Rupture: Range 1.5 to 4.5, Average 2.34
6. Pullout Resistance: Range 1.5 to 2.0, Average 1. 68
Although in several cases the factors of safety are
substantially less than those normally recommended for
external stability consideration, the adequacy of the design
cannot be determined without studying the actual design
procedure since other conservative assumptions may have been
made.

3.5 Project Costs

Table No. 7 presents the reported cost data for a number of

the projects. The total costs include fabric, backfill and
equipment and labor costs per square meter of wall face. The
costs range from $39/m 2 to $450/m 2 . The average cost of the
permanent walls was $158/m2, not including project numbers 4,
5 and 6 for the reasons mentioned below and project number 8
since tire reinforcement was utilized. The average cost of
the , temporary walls was $83/m 2 •

The lower bound cost is for the SR 522 project in washing~on

[No. 241. Five temporary walls were constructed USlng
UV-stabilized geotextiles, which resulted in a cost savings
 269

since facing materials were not required. The New York Route
22 project [No.5) was constructed at a cost of $450/m 2 • The
high cost resulted from the use of expensive crushed stone
backfill ($24/cy). Chassie (1984) noted that the high costs
of Projects No. 4 and 6 may have resulted from the walls
being changed ordered into the project.

TABLE NO.7: PROJECT COSTS

Project Total Project Cost

Number ($/m 2 of wall face)
------------------------------~~

1 129
2 127
4 320
5 450
6 312
7 203
8 270
9 134
13 255
15 92
23 46
24 39
27 129
29 96
30 160
32 65
33 109

Figure number 1 compares the cost of various retaining wall

systems, including the average costs for the projects
surveyed. The cost dati for projects 1, 7, 9, 13 and 27 was
superimposed on the cost comparison chart [33). For walls
less than 30 ft in height, geosynthetic reinforced walls
would appear to be more economical than metallic reinforced
soil walls.

4. TECHNICAL ASSESSMENT OF MONITORED PROJECTS

Of the projects reviewed, thirteen were instrumented to

varying degrees beyond simply taking surface survey
measurements. As previously indicated, Table 4 summarizes
the instrumentation program for those projects. The table
was set up to include the full compliment of instrumentation
that might be considered on a project to assess the principal
design assumptions, the principal design considerations
include:

1. The location of the potential failure surface

corresponding to the location of maximum stresses
within the reinforcement.
270

Q)

~
~ ~60~-------r------~--------~------~
~
~ I reinforced concrete . /
Q) I retaining wall •
S 460 t------------:r-----~-___i
~
C'Q
/ /'
;::j / ~ reinforced concrete
0" / ,
If.l 360 ~_______~_....,..,01:..---,......_ _
cr_ib_w_a_ll_ _ _--1
~ /~./'

~
0..
~
........
~""",.,.
~ • ",.,. metal crib wall

... ....
260 '"""'-.........r::;....._.....r:::;...._ _ _ _ _ _ _ _~,...._I.....,:..-.---~
""0
o ••
• •metallic reinforced soil wall
.....=
..... 160 t----~'ZfIIII""C-----------___I
Cd
~
C+-t
o 60~------~------~------~------~
~
If.l
oCo) 3 6 9 12 I~
height of wall in meters

FIGURE 1 Derived from a comparison prepared by the

California Department of Transportation in
cooperation with the Federal Highway Admin-
istration, this chart compares the -materials
and backfill costs of competing systems.
(adapted from vsl corp., 1981)
 271

2. The stress level resisted by the reinforcement as

determined from lateral earth pressure wi thin the
reinforced mass based on the vertical pressure at
critical levels.

3. Both short term and long-term horizontal

displacements.
4. External structural stability including the
magnitude of vertical displacement and the
foundation support stability.

As indicated in Table 4, most of the project instrumentation

programs concentrated on lateral movement and overall
settlement. vertical movement was typically measured by
external surveys and internal measurements using varying
combinations of settlement plates and anchors, vertically
spaced inductance coils, and horizontal inclinometers.
Lateral deformation was usually measured with some form of
extensometer and/or inclinometers. In many cases the results
of such measurements were found inconclusive as the
instruments were not sensitive enough to measure movements
that were usually smaller than anticipated.
A full compliment of instrumentation, including the
measurement of strains in the reinforcement and soil
pressures in the reinforced soil mass was noted for eight of
the projects (Numbers 3, 9, 18, 19, 20, 26 and 38). Project
No. 35 was also fully instrumented; however, results were not
available at the time of this writing. Due to significant
variations in the structures instrumented, including the type
of backfill, reinforcement, facing panels and foundation
conditions, it is very difficult, if not impossbile, to
develop definitive conclusions from the instrumentation
results of the projects surveyed. Therefore, the following
includes a synopsis of the results rather than a comparison
of the results.
In the waterways Experiment Station Project [3], a failure of
the polymer strip type reinforcing was ' observed.
Observations indicated the failure may have been due to
excessive deformation of the reinforcement near the face of
the wall and possibly at facing connections. Pressure cells
embedded in the soil near the face of the wall indicated an
increase in pressure with increased overburden as shown in
Figure 2. However, the measured lateral pressures were less
than those predicted by Rankine theory. The pressures
increased significantly as the wall started to fail. The
pressure measurements indicated that the reinforcement became
more effective with increased lateral movement. As the
reinforcement started to yield, the lateral stress was
apparently then transferred to the wall face.
272

theoretical Ka line
O"h = Ka = tan 2 (45 - </J/ 2 )
7.0 O"V
,--.., /
cd </J = 36°
~
~
'--"
6.0 I
......
...... I
I
CI)
Co)

~ ~.O I
;:j
rn I
rn
CI) I
~ 4.0 I
I
S
o I
~
"'0 3.0 I
CI)
I
~
;:j I
~ 2.0 I
S Legend
...c= D--O Cell 81H a.3m above bottom
b 1.0 0--0 Cell 85H 1.5m above bottom
tr---6 Cell 89H 2.7m above bottom

20 30 40 ~o 60
O"v calculated from overburden (kPa) ,

FIGURE 2 Comparison between the measured horizontal

stress and the calculated vertical stress
for wall reinforced with rubber-coated
strips.
(Case No.3)
 273

For the Glenwood Canyon Walls [9], very small strains were
measured in high elongation needle punched non-woven fabrics,
even though the design methodology would indicated factors of
safety near 1. The results were attributed to the apparent
improvement in stress-strain properties of the geotextiles
under the confined soil condition. This project emphasizes
the need to use confined stress-strain values in predicting
performance of structures constructed with non-woven
geotextiles.

As previously indicated, in the Tarque Verde Project [18],

full-height tilt up concrete facing panels were used. Strain
measurements on the geogrid reinforcement indicated that the
stress levels in the reinforcement were less than
anticipated. The location of maximum reinforcement stress
was found to be 18 to 19 degrees from the vertical, - as
measured using bonded resistant strain gages and 24 to 25
degrees from the vertical (or approaching a Rankine surface)
as measured with inductance coils. The researchers felt that
the bonded resistant strain gages provided the most reliable
data. The location of maximum tensile strains at a steeper
surface than predicted by Rankine would tend to agree with
model studies of full height facing panels (Christopher,
1987) . As shown in Figure 3, lateral stress measurements
using soil embedment cells near the face of the wall found
low stress levels, again indicating that the reinforcing was
adequately resisting lateral stress within the reinforced
soil mass. Higher lateral stress levels were observed near
the base of the wall which may have been due to embedment of
the face at the base and connection of the facing panel with
an embedded footing. As the face was not free to move at the
base, the reinforcing could not sufficiently deform to carry
significant stress. Vertical soil stresses measured wi th
embedded pressure cells found greater pressures at the base
of the wall than predicted by trapezoidal distribution,
overburden stress calculations or Meyerhof distribution
(Figure 4). However, the stress levels were surprising lower
near the face of the wall. This was attributed to wall
friction and to soil arching over the load cells. ,
The Lithonia Project [19] was similar to the Tarque Verde
Project, except that articulating facing panels were
utilized. A good summary of the two projects was provided by
Berg, et al (1986). The results were less conclusive than
the Tarque Verde Project, however, again lower stress levels
were measured near the face of the wall (Figure 5) again
indicating resistance to lateral stresses by the
reinforcement. The vertical pressures measured below the
base of the wall (Figure 6) appeared to be less than those
measured at the Tarque Verde project, however, data was
limited to only two locations.

Probably the most extensively instrumented wall was the Gaspe

Peninsula Reinforced Soil Seawall Project [20]. Strain
measurements indicated that the line of maximum tension in
274

lateral pressure O"h (kPa)

10 20 30 40

Tolerances
soil weight ± 6%
field measurement ± 20%
\
~
~ field measurements

~.
~,
\<" r trapezoidal distribution
\ .'(
.
,"
\

lateral pressure l' ". Meytrhof type distribution

FIGURE 3 Comparison of Measured Horizontal Soil

Pressures to Predicted Values - Tucson
Wall (Case No. 18)

horizontal distance behind wall (m)

............ 2
cd

f.
~
..!c:I trapezoidal distribution
"--'
field measurements
> .'
b
Q) ..,...,' overburden stress bh)
'rn""
:::l
rn
~eyerhOf
Q)
- -[
'""
0..
....... 80 Tolerances type distribution
....
cd
u
-+-"
soil weight ± 6%
field measurement ± 20%
'""
Q)
:>
FIGURE 4 Comparison of Measured Vertical Soil
Pressures to Predicted Values - Tucson
Wall (Case No. 18)
 275

lateral pressure O"h (kPa)

,--..
10 20 30 40
S
'--' Tolerances
.......
....... soil weight ± 6%
Q)
u field measurement ± 20%
.....~--field measurements
~
o
.......

II.Y---
\ .
\, . trapezoidal distribution

" .
\
\

IRankilne
atera pressure
~ Meyerhof type distribution

FIGURE 5 Comparison of Measured Horizontal Soil

Pressures to Predicted Values - Lithonia
Wall (Case No. 19)

horizontal distance behind wall (m)

2 3 4

V fi,ld m'M""m~'"

~ / trapezoidal distribution

./. overburden stress (,h)

/
.......
cd
.....u
---r.-~--
/
--
' - Meyerhof type distribution
~
~
Q)
./. Tolerances
> 200 soil weight ± 6%
field measurement ± 20%
FIGURE 6 Comparison of Measured Vertical Soil
Pressure to Predicted Values - Lithonia
Wall (Case No. 19)
276

the reinforcement was very close to a Rankine surface

(Figures 7 and 8). The measurements indicated that the
tensile loads in the geogrid reinforcement was below
tolerable levels. Future measurements at this project will
be of significant interest as the polymer reinforcement is
subjected to cyclic loading.

The Cascade project [26] is significant in that it was

exposed to very high temporary surcharge loads. In all of
the previous projects, it was reported that the stress in the
reinforcement was less than that predicted by the design
model. For this pro~ect, the instrumentation results would
indicate that stress levels were very close to those
predicted and, during maximum surcharge loading, were
slightly higher than anticipated . The locus of maximum
stresses obtained from strain gage measurements (Figure 9)
found the highest stress levels at the top of the wall
towards the back of the surcharge loads, similar to that
observed by Wichter et aI, 1986.

It is very apparent that more controlled research test walls

are needed to evaluate design methodologies and develop
predicted capabilities. Even so, for simple structures (i.e.
walls less than 30 ft height with low surcharge loading,
granular backfill, and facing elements that are free to move)
current design methods using unconfined stress-strain data
appear conservative and are thus adequate. However, for
continued development and better predictions of performance,
an extensive amount of research is still needed.
5. CONCLUSIONS

Polymeric reinforcement is gaining wide spread use and is

helping to solve many problems that could not formerly be
solved using rigid retaining structures. It is estimated
that over 200 structures have been constructed in North
America. Applications of polymeric reinforcement range from
support of a bicycle path over peat deposits in Anchorage,
Alaska to protecting a sea coast in Quebec, Canada.
Innovative uses of these materials can help solve
construction problems, such as the Cascade Michigan Dam
project where a geogrid wall was used to widen a dam and
supported a 840 kN crane with a 100 kN lifting capacity for a
total surcharge of 200 kN/m2 during lifting.

Polymerically reinforced retaining structures have also been

shown to be very cost effective alternatives to conventional
retaining structures. Temporary walls have been constructed
at costs as low as $39/m 2 of wall face. Costs for permanent
walls range from $130 to $260/m 2 of wall face.

To ~he author's knowledge, all of the projects surveyed have

performed satisfactorily in serving their intended purpose.
Unfortunately, problems associated with construction or
 277

j.,.o", ...j
I

, .' .'
..
I

T I
1.0",
...L I,~"mrmd Lin, of
t/ Mr.tmurn Tension
I
I •

III III, ,,
~ Ellt.MDm.f.r
(a). Section 1

--
-+- Clamp L()fld Ce 1/

+ Totol S'ren C,II

Sfrt1ln Gag,

• A".rom,f,r
Tl/erm()C(Jupl,
Obs.rv.d Loc~on of
Mf11timvm Strain

~Om~

,
I
I
I

T 1.1
I.Om
,'1
...l , -ESlimalld LIn, of
I MaX/lflii'" Tension

;'

(b). Section 2

FIGURE 7 Instrumentation Layouts (Case No. 20)

278

0.8

.... 0.7
~
~ 0.6

II)
...
~
o.e
I.-
0.4
~
~
~
0.3

~ 0.2
~
~ 0.1

0
1.0 I.e 2.0 2.e 3.0. 3.e 4.0 4!5
DISTANCE BEHINO WALL FACE (m)

(a). - Geogrid Strain, Section 2, Level 3

..... 004
!
~
~ 003
~
"t ·EIII,nSDm.'",
~ Results
~

_-----0
t:) 002
-.J
~
~
~ 0.01
<t
C
:t:

\
1.0 1.5 2.0 2.5 30 3.5 4.0 4!5
DISTANCE BEHIND WALL FACE (m)

(b). - Wall 1'1.11 Strain, Section 2, Level 3

DIll. of FIwoding
+ - - - t 85/08/01 • Fill ClfJ" to Boffo", of 7bp PaHI (EI. 2.8)
o c 85/08/06 Fill III T(lp (If Wall (£l4.:J)
6 r 85/10/:1/ Fill ftJ R(ItIt/ Elwaf/(lII (£1.5.1)
- - - . 86/06/1/ R.fJding AfI.,. Finl Winl.,

FIGURE 8 Strain Measurements (Case No. 20)

 279

10000 10000
~
9000 9000
E 8000 8000
'----
.•" El'dolCO'lCtn.l:._
....... ftlr2'~
7000
E . _ ... 1ter~moncIIs 7000
:::L 6000 "'AftM_~
~""-An. 21 .,.. (~ ~) 6000
c 5000 5000
LEVEL 8 .~
4000 4000
U;
0 3000 3000
,-
u 2000 2000
~ 1000 1000
o~-.--.--.--.--.~,--,~,
0.3 0.6 0.9 1.2 1.S 1.8 2.1 2 "4
10000 - - ElWSoIco.lructioo'l
"·An.r2t~

9000 ·...·Aftr'..-.
E 8000
- . Alt.. aa-.
"." An.- 21 <iQy.I.
Io4dinq
(moqne~( .. a~)
'----
E 7000
.......,
:::L 6000
c 5000
LEVEL 6 "0 .------. '.
b 4000
Ul
3000 '\....,\
0
'-
u 2000
~ 1000
0
2" 4

d.
10000 10000
~
9000 -£ftIIII«~_
_. MtM21
~OOO
.... End ot ccn.tn.octiGn
E 8000 .... Nt. !I mon'h 8000 ....... An.. 21 4ay.I

'----
_ After 5 month,
..... Nt¥~'ClGall'l<J
7000
E 7000 .... £nd ct con:rtruc:.tictI (~Iic
..... NW 21 dG')'ll (~tic: q09"~)
9C'9G)
•
.......,
::l. 6000 6000
c 5000 50PO
LEVEL 3 "0
'- 4000 4000
U; 3000
0 3000
'-
u 2000 2000
~ 1000 1000
o. O~~--.-~--'--'---'-'r-­
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 o 0.3 0.6 0.9 t .2 1.5 1.8 2.1 2.4
Position from f(lee (m) Position from face (m)

FIGURE 9 Strain Distribution Along Reinforcement

During Monitoring Events (Case No. 26)
280

performance were not reported for the projects reviewed.

Information concerning problems would have provided valuable
information for future users.
The authors will continue to update the summary tables as
information on new projects and additional information on
existing projects become available. Any help from the
readers in this effort would be appreciated.
ACKNOWLEDGEMENTS

The authors wIsh to thank all of the individuals who provided

p-roject information and assistance in preparation of the
summary tables, Terry Harrison for typing the manuscript,
Maria Flessas for reviewing the manuscript and Horw·ard Faye
for computerization of the summary tables.
 281

REFERENCES

1. Bell, J.R., and Steward, J.E., "Construction and

Observations of Fabric Retained Soil Walls,"
Proceedings, International Conference on the Use of
Fabrics in Geotechnics, Paris, France, Vol. 1, April
1977, pp. 123-128.
2. Bell, J.R., A.N. Stilley and B. Vandre, "Fabric Retained
Earth Walls," Proceedings, 13th Annual Geology and soils
Engineering Symposium, Moscow, Idaho, 1975, pp. 271-287.
3. Steward, J.E. and J. Mohney, "Trial Use, Results and
Experience Using Geotextiles for Low-Volume Forest
Roads," Proceedings, 2nd International Conference .on
Geotextiles, Las Vegas, Nevada, Vol. 2, August 1982,
pp.335-340.
4. Mohney, J., "Fabric Retaining Wall, Olympic National
Forest," Highway Focus, U.S. DOT, Federal Highway
Administration, Vol. 9, No.1, May 1977, pp. 88-103.
5. Al-Hussaini, M.M., "Field Experiment of Fabric
Reinforced Earth Wall," Proceedings, International
Conference on the Use of Fabrics in Geotechnics, Paris,
France, Vol. 1, April 1977, pp. 119-121.

6. Al-Hussaini, M.M. and E.B. Perry, "Effect of Horj.zontal

Reinforcement on Stability of Earth Masses, "Technical
Report No. S-76-11, 1976, U. S. Army Engineer Waterways
Experiment Station, Vicksburg, Mississippi.
7. Douglas, G.E., "Design and Construction of Fabric
Reinforced Retaining Walls by New York State,"
Transportation Research Board, Transportation Research
Record 872, washington, D.C., 1982, pp. 32-37.
8. Chassie, R.G., "Geotextile Retaining Walls, Some Case
History Examples," prepared for presentation at the 1984
NW Roads and Streets Conference, Corvallis, Oregon,
February.
9. Walkinshaw, J.L. and K.A. Jackura, "New Developments in
Retaining Wall Design 38th California Transportation and
Public Works Conference," Oakland, California, May 1986.
10. Bell, J.R., R.K. Barrett and A.C. Ruckman, "Geotextile
Earth-Reinforced Retaining Wall Tests: Glenwood Canyon,
Coloardo," Transportation Research Board, Transportation
Research Record 916, washington D.C., 1983, pp. 59-69.
11. Barrett, R.K., "Geotextiles in Earth Reinforcement,"
Geotechnical Fabrics Report, March/April 1985, pp.
15-19.
282

12. Busbridge, J.R., "Stabilization of CP Rail Slip at

waterdown, Ontario, Using Tensar Grid," Proceedings of
the Symposium on Polymer Grid Reinforcement in Civil
Engineering, London, U.K., 1984, paper No. 2.3.
13. Fluet, J.E., "Geogrids Enhance Track Stability," Railway
Track and Structures, June 1984.
14. Tensar Corporation, "Case Study: CP Rail Slope Repair,"
Morrow, Georgia.
15. Tensar Corporation, "Case Study: U.S. 69 Slope Repair,"
Morrow, Georgia.

16. Bell, J.R., T. Szymoniak and G.R. Thommen, Construction

of a Steep Geogrid Retaining Wall for an Oregon Costal
Highway, Proceedings of the Sympsium on Polymer Grid
Reinforcement in Civil Engineering, London, U.K., 1984,
Paper No. 6.3.
17. Szymoniak, T., et aI, "A Geogrid-Reinforced Soil Wall
for Landslide Correction on the Oregon Coast,"
Transportation Research Board, Transportation Research
Board 965, washington, D.C. 1984, pp. 47-55.
18. Bonaparte, R. and E. Margason, "Repair of Landslides in
San Francisco Bay Area, "Proceedings of the Symposium on
Polymer Grid Reinforcement in Civil Engineering, London,
U.K., 1984, Paper No. 2.4.
19. Forsyth, R.A. and D.A. Bieber, "La Honda Slope Repair
with Geogrid Reinforcement," Proceedings of the
Symposium on Grid Reinforcement in Civil Engineering,
London, U.K., 1984, Paper No. 2.2.
20. Berg, R.R., R. Bonaparte, R.P. Anderson and V. E.
Chourey, "Design, Construction and Performance of Two
Geogrid Reinforced Soil Retaining Walls," Proceedings,
Third International Conference on Geotextiles, Vienna,
Austria, 1986, pp. 401-406.
21. Anderson, R.P., "Soil Reinforcement Objective: Polymer
Geogrid Replaces Galvanized Metal Strips in
Concrete-Faced Retaining Walls," Geotechnical Fabrics
Report, January/February, 1986.
22. "Geogrids Used to Reinforce Retaining Walls," Rocky
Mountain Construction, May 1985.
23. "Jasper Construction Builds Retaining Wall Utilizing
Tensar Geogrid as Reinforcement," Dixie Contractor,
October 1985.
 283

24. Berg, R.R., P. LaRochelle, R. Bonaparte and L. Tanguay,

"Gaspe' Peninsula Reinforced Soil Seawall Case History,"
Proceedings, ASCE Symposium on Soil Improvement,
Atlantic City, New Jersey, April 1987.
25. Christopher, B.R., "Geogrid Reinforced Soil Retaining
Wall to Widen an Earth Dam and Support High Live Loads,"
Proceedings, Geosynthetics '87 Conference, New Orelans,
Louisiana, February 1987.
26. Scott, J.D., et aI, "Design of the Devon Geogrid Test
Fill," Proceedings, Geosynthetics '87 Confernece, New
Orleans, Louisiana, February 1987.
27. "Class A Prediction Exercise for Reinforced Earth
Walls," NATO Advanced Research Workshop to be held June,
1987, Royal Military College of Canada, Bulletin No.1.
28. Christopher, B.R., unpublished STS Consultants, Ltd.
Report on Task C Field Test Wall Construction - Prepared
for FHWA, washington, D.C., Contract No.
DTFH61-84-C-00073, "Behavior of Reinforced Soil", 1987.
29. Castelli, R. and G. Munfakh, "Geotextile Walls in
Mountainous Terrain," Proceedings, Thi rd International
Conference on Geotextiels, Vienna, Austria, 1986,
pp.459-463.
30. Christopher, B.R. and R.D. Holtz, "Geotextiles
Engineering Manual," Prepared for FHWA, National Highway
Institute, Washington, D.C., Contract No.
DTFH61-80-C-00094, 1985.
31. Koerner, R.M., "Designing with Geosynthetics," Prentice
Hall, Englewood, Cliffs, New Jersey, 1986.
32. Bell, J.R., "Design Criteria for Selected Geotextile
Installations," Proceedings, First Canadian Symposium on
Geotextiles, Calgary, Alberta, 1980.
33. Holtz, R.D. and B.B. Broms, "Walls Reinforced by Fabrics
Results of Model Test," Proceedings, International
Conference on the Use of Fabrics in Geotechnis, Paris,
France, Vol. 1, April 1977, pp. 113-117 .
34. VSL Corporation, "VSL Retained Earth, Technical Data,"
Rock and Soi 1 Stabilization Systems, Los Gratos,
California, 1981.

35. Wichter, L. Risseeuw, P. and Gay, G., Grossversuch zum

Tragverhalten einer Steilwand aus Gewebe und Mergel.
proceedings of the Third International Conference on
Geotextiles, Vienna, Austria, 1986.

36. Personal Communication, Ryan R. Berg, The Tensar Corp.,

1987.
GEOTEXTILE-REINFORCED RETAINING STRUCTURES
A FEW INSTRUMENTED EXAMPLES

Ph. DELMAS
Laboratoire Central des Ponts et Chaussees, France
J.C. BLIVET
Laboratoire Regional des Ponts et Chaussees de Rouen, France
Y. MATI CHARD
Laboratoire Regional des Ponts et Chaussees de Nancy, France

1 - INTRODUCTION

The principle of retaining structures consisting of a

reinforced soil mass has led to routine applications primarily
through the work of la Terre Armee (TA). From this viewpoint,
geotextile-reinforced structures may be compared to TA
reinforced-earth structures.

These two processes do have a number of points in common :

the retaining function is performed by a gravi ty structure of
reinforced soil; reinforcement is by elements in which
tensile stresses predominate, placed horizonta~ly when the
fill is laid down; and the facing is generally vertical.

But the similarities between the two types of structure end

here, and an analysis of their internal behaviour reveals
differences.

In TA reinforced-earth structures, the shape of the rein-

forcing strips serves to mobilize the dilatancy of the soil at
the level of local soil-reinforcement interactions (in the
pull-out mode). The resulting additional local vertical stress
at the reinforcement makes it possible to mobilize substantial
friction in spite of the narrowness of the reinforcements,
especially in the upper part of the wall where the vertical
stresses are smallest. But this beneficial working mode calls
for the use of soils that are dilatant, and so suitably
prepared (Schlosser and Guilloux, 1981).
285

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 285-311.
© 1988 by Kluwer Academic Publishers.
286

Geotextiles are in sheet form, and because of their two-

dimensional character do not generally make use of the
dilatancy of the soil. It should, howerer, be noted that some
geogrids can bring out the dilatant behaviour of the soil. But,
generally speaking, because of the large contact area between
the soil and the geotextile (figure 1), the additional normal
stress that serves to mobilize dilatancy on strip
reinforcements is not needed to mobilize the friction
necessary for stability. With geotextiles, then, it will be
possible to use non-dili tant soils, and in particular any
material likely to be used as fill.

As for the tensile strength of the reinforcements, table 1

gives comparative strength values for one layer of reinforce-
ments and a width of one metre. It can be seen that the two
processes yield comparable values.

The stiffness of the reinforcements, again calculated for

one layer and a width of one metre, is between 25 and 100
times smaller in the case of geotextiles. This results in
differences in behaviour, character i zed by larger strains
within the geotextile-reinforced structure for similar forces.

Figure 1 - Comparison of the friction behaviour of "Terre

Armee" reinforcement and geotextiles

These findings explain why, in observations of geotextile-

reinforced structures, an effort has been made to measure the
strains of the reinforcements and of the soil. In the case of
real structures, it is important to be able to make sure of a
level of strain that is acceptable for the structure, whether
locally, in the soil (to remain below the peak shear strength
values), in the geotextile or at the contact surface, or more
generally with respect to the superstructures or the facing.
 287

Table 1 - Comparison of stiffness and strength values of TA

reinforced-earth reinforcements and geotextiles
(one layer one metre wide)

"Terre annee n
Geotextile
(b = 40 rrUll, spacing 1,20 m)

Maximum tensile
57,8 kN/m 50 ~ 300 kN/m
strengh

~llowable tensile
atrengh. 26,9 kN/m 5 /"'- 60 kN/m
(permanent use)

stiffness 28 000 kN/m 500 _ 2000 kN/m

Based on a few examples of instrumented real structures, we

shall attempt in what follows to determine the influence on the
behaviour of the structures of

- the method of placement of the formwork and of the facing

- the use of clayey soil ;
- the long-term behaviour of the geotextiles.

For practical reasons, this analysis will be limited to the

investigation of structures with flexible facings :

2 - INFLUENCE OF PLACEMENT METHOD ON THE BEHAVIOUR OF

GEOTEXTILE-REINFORCED STRUCTURES

In practice, the stabi 1 i ty of the structures with respect

to the equilibrium of forces is ensured by the tensioning of
the reififorcement sheets inside the soil mass. However, the
overall behaviour of the structure will depend primarily on how
the sheets of geotextile are tensioned and how the strains are
induced. Indeed, for a given structure, overall static
equilibrium can be ensured for different stress conditions in
the geotextiles depending on how the strains occur in the soil
mass.
288

Accordingly, placement will undeniably be a key stage in

the generation of the forces in the reinforcement sheets. The
few instrumented structures described here show, in effect,
that in certain cases the final forces can be reached as soon
as the construction of the structure is completed, independen-
tly of the of the subsequent loading conditions.

This is explained by the stiffness of the geotextiles,

which means that rather large strains of the soil mass are
required to generate the tensile forces that ensure stability.

It follows that correct placement of the sheet is critical:

no pleats, if possible manual tensioning before compaction of
the fill, etc. However, the type of compaction and the way in
which the fill is cased will in general be the two factors
that determine the final strains in the soil mass. In addition,
the facing, and its relative stiffness, will also be a factor
to be taken into account in understanding the limit-state
strains of the structure.

As an example, we may mention the first geotextile-reinfor-

ced structure built. It is in many respects a reference
structure that sheds some light on the behaviour of geotex-
tiles used for reinforcement.

2-1 Embankment of A1S motorway

This structure, built on an experimental basis in 1971

(Puig and Blivet, 1973), is one of the first geotextile-
reinforced retaining structures built with a clayey soil. At
the site of the structure, the embankment of the A1S motor way
is in a compressible valley with, under the s~rface layer of
old fill, a 3-m layer of peat having the following mean proper-
ties w = 300 % ; C = 3.2 ; e = 6.8. The fill material is a
clayey sand-gravel consisting of a mixture of weathered chalk,
clay, and flint. The water content of the fraction under 20 mm
ranged from 16 to 21 % at placement, corresponding to a satura-
tion Sr of 94 to 98 % (at the normal Proctor
optimum, w = 13 %, d = 19.0 kN/m3). This explains, in
particular, the rubber cushion effect observed when the emban-
kment was built up.

Figure 2 shows the standard cross-section of the structure,

reinforced with layers of needle-punched polyester nonwoven
(Bidim U34) every SOcm

The form used was a temporary bank having a batter of 3 in

2 (figure 2).
 289

Seven month after the facing was uncovered, the structure was
loaded with an embankment 4 m high having a batter of 3 in 2.

- st:r-ain gauge
.'- ......... _,"'".,- settlement meter

......
~
:
~,
,
clayey ........ I..·~"·"~r r!J level bench mark
gravel .......' ..... "'" -iilr''''''''''''-"'t~ "-

, .." ..n.. ""Iii"~.." ..~..t, ""

.. "
_ ..,......
"'""-...........-...- "liII--............~ "
.......... ........ .. .......... - " ,
...... - ........ - -'
.. ~~~~~~~~~~~. . . .L-________________~"~
o .... <:.. " p ~. 0 ." • V" ,
.. . . , . . . o· '. :

old fill

peat

Figure 2 - AlS motorway detail of construction with

temporary bank and arrangement of measuring ins-
truments

This experimental embankment, the total length of wich was 20 m,

was instrumented with :

- settlement meters placed in the base of the embankment;

- strain gauges placed on the reinforcements (these gauges
were made of resistive constantan wire mounted on a plastic
support ; they had a useful length of 6 cm and were attached
to the geotextile by clamping their ends between two steel
plates)
- an inclinometer tube placed in the body of the embankment
- level bench marks and plates on the facing.
290

Photo 1 - A15 motorway overall view of completed struc-

tu.re

For all observations, the reference measurement was made

on 23 March 1971 just after the wall was uncovered (photo 1).

Immediately after the removal of the "formwork" fill,

measurements of the horizontal displacement of the facing
revealed a displacement of 15 td 25 mm at the top of the
structure and a displacement of 6 mm at the toe, so indicating
tilting of the top.

After the top of the embankment was loaded, the measure-

ments revealed an overall displacement of about 20 mm. These
displacement values are also confirmed by the inclinometer tube
in the soil mass.

It should be noted that the interpretation of these

measurements is complicated by the presence of substantial
settlements of the foundation soil : the total settlement, one
month after loading, reached 1.10 m, with a maximum differen-
tial settlement of 0.25 m in 3 m

Eyen though the strain gauges on the geotextiles did not

survive long (of the 35 placed, only 6 were still functioning
three months after the removal of the "formwork" fill), they
gave some idea of the strains and forces in the geotextile
sheets. The following may be noted :

- in general, the strain of the geotextile increased as the

distance from the facing decreased ;
 291

- the maximum values measured (before surcharging) reached

1.7 to 1.8 % in the geotextiles at the toe of the structure.

In interpreting these findings, one must first of all bear

in mind the very special mode of placement. Using an adjacent
bank as a form prevents strain of the geotextiles before the
removal of the form; again, especially near the facing, the
loss of compaction caused by the process interferes with the
proper confinement of the end of the layer of material
(photo 1). Moreover, the removal procedure used causes the
wall to work as a trench retaining structure, by stressing
first the uppermost layers of reinforcements, then the under-
lying layers, as evidenced by the tilting of the top when the
form was removed.

It is reasonable to suppose that the tensile forces engen-

dered in the reinforcements depend to a large extent on this
particular stressing mode. Nor can it be ruled out that the
lowest layers were stressed at the back of the soil mass to
ensure the overall stability of the structure when the forms
were removed.

So far, there are few instrumented works built using global

forms, in all likelyhood because of the large force required
for the temporary stability of the formwork. But the example of
the "La Houpette" structure on National Highway 4, built in
France in 1986 using this process, should be mentioned.

2.2 - La Houpette embankment

To carry out the widening of a road on an embankment 4m

high and 300 m long, it was decided to build a' geotextile-
reinforced structure consisting of a vertical part 1.4 m high
reinforced by two layers of ueo 44614 topped by a bank having
a batter of 3 in 2 (figure 3). For practical reasons, the
contractor chose to build the lower part in forms placed all at
once.

The fill is anggreqate having a continuous 0/250-mm gra-

ding and the following geotechnical properties : rd = 22 kN/m 3
c' = 0 kPa ; lj" 3 5 0 to 4 0 0 •

Because of the size and rather sharp angles of the grains

of the fill material, a test of compaction on geotextile was
carried out. Figure 4 gives the tensile curves of the geotex-
tile before and after this compaction. It can be seen that, in
this particular case, the compaction had relatively little
effect on the stress/strain curve of the product.
292

However, in the course of specific tests carried out on

various products placed between two layers of flint-bearing
clay dumped from a height of 1.5 m and compacted to 95 % of the
normal Proctor optimum in layers 30 cm thick, Perrier (1986)
found that the compaction could decrease the maximum tensile
strength of the geotextile by as much as 30 %, and alter its
stiffness, either increasing or decreasing it (table 2).

/-7- Highway

T 3.'0 -m.. /
lV
3 /

/
Ct
,/ 0 . 1
,/ /

././ """:•.:::- 0 _ _ 0 , / /

---'
0.70 ......

\~~~\~~'~'i~~~;;;';;-/-"-:;"':';'... . . . . ,. . ." 0
• strain h 1.4 m
€ ~ . .".. 2 ~ 4.0 m
. o h
!t

0.70"""
,
t
\/
~--;-;::~ 3.5'0 __

4.0(/
I

Figure 3 - National Highway 4 : standard cross section

and arrangement of measuring instruments

This embankment was instrumented with cable-type strain

measuring devices. Each measuring point consists of two cables
attached to the geotextile 20 cm apart (in the direction of the
stress), protected from the friction of the fill material by
a sheath, that extend out through the front of the soil mass.
The relative displacement of the two cables is measured with
respect to a fixed reference about 1 m from the facing, with
the same tension applied to both. This procedure has the advan-
tage of being easy to implement and relatively inexpensive. It
makes it possible to estimate strains that exceed 0.5 % and
also, provided that the measurement bench mark is levelled,
makes it possible to determine the absolute displacement of the
points of the soil mass.
 293

Table 2 - Influence of compaction on the tensile properties

of the geotextile (after Perrier, 1986)

BeJ:ore compaction Loose after compaction

a - a
-€ -€
-a (kN!m)
-€ (%) 0
(%)
0
-
0 0 cr0 €
0
(%)

· NTA PES 41,8 51,2 8,1 21,9

·t t PES ST 158,6 7,6 35,2 18,4

· PES AJ 115,3 7,3 30,2 15,1

·t t PP X 40,9 15,2 15,4 5,9

·t PP N 96,4 8,7 8,7 0,0

· PET 55,0 22,5 21,5 19,1

· G PET 19,1 7,3 0,0 26,0

0(

(leN/",)
0(,= 12.0 IeN/",

-100
£r = 1t r.

Soil: clayey gravel

So D = 250 mm
before compaction
af~er compaction

o
o s 40

Figure 4 - National Highway 4 : comparison of tensile stress/

strain curves of the woven polypropylene
geotextile before and after compaction
294

The deformations measured immediately after the removal of

the forms from the two layers, before the upper embankment was
built, and at the end of its construction, are given in table 3

Table 3 - National Highway 4 : deformations measured in

geotextiles at measurement profile no. 2

Facing'
Point of measurement 1 2 3 4

Height of embankment 1.4 4.0 1.4 4.0 1.4 400 1.4 4.0
(m)

Geotextile 1 0 0 0 0.5
1.0 o 5 0 0.5

Geotextile 2 0 0 0 005
2.0 0,5 0.5 0,5

Otherwise, the absolute displacements recorded revealed,

after the removal of the forms, an overall displacement of the
sheets of geotextile 7 mm for the bottom sheet and 5 mm for
the top sheet. This displacement was simultaneous with the
strain of points no. 4 of both sheets, near the facing.

It will be noted that, as regards the strains of the

sheets, the form removal stage corresponds with a mobilization
along a surface of maximum tension located near the facing,
while the stage of embankment construction mobilizes the rein-
forcements farther back in the soil mass, so justifying the
anchorage lengths determined in the preliminary design stage.

From these findings, we may conclude that the rather rudi-

mentary fill placement method used at this site (photo 2) is
doubtless sufficient to explain the overall displacements
observed when the forms were removed. It would seem that no
special care was taken with the compaction in the vicini ty of
the facing.
 295

It would also seem that the principle of a single form for

the whole facing leads, when the form is removed, to a mobili-
zation of strains in the sheets in accordance with those
anticipated in the preliminary design calculations. Having said
this, we should point out that the final distribution of
strains in the geotextiles in fact corresponds to the sum of
the strains resulting from each of the two stages of
constructions.

Photo 2 - National Highway 4 placement of fill

Generally speaking, geotextile-reinforced structures are

built with partial forms for each layer. In what follows, we
shall not analyze structures built using forma supported only
by the next lower layer, even though in historical fact this
was one of the first processes used. This is because the use
of a form supported by the next lower layer assumes that there
is no deformation of this layer when its form is removed, or at
least that any deformation is uniform, which js practically
never th,e case.

Obtaining a vertical facing calls for a series of adjuste-

ments, all the more delicate in that a deviation from the
profile caused by an outward displacement of the form cannot
be cor rected, as shown by the exa mp le of the CER' s
296

experimental structure at Rouen, the defects of which were

aggravated by the use of very moist silty materials (photo 3).
In short, this process is not a very reliable way of building
structures with vertical facings.

We shall accordingly restrict ourselves to analyzing the

deformations engendered by the process in which the form for
each layer is placed directly above a fixed point and supported
by a fixed reference, derived from the patent held by the
Laboratoires des Ponts et Chaussees (1985) and worked by the
MUR EBAL company.

In this connection, we may mention the Langres struGture.

Photo 3 - Experimental structure bUllt by CER of Rouen :

the difficulty of obtaining a ver~cal facing
using a form resting on the next lower layer

2.3 - Langres structure

Where it goes round the ramparts of the town of Langres,

national highway 19 is on the uphill side of a retaining wall
4 m high and includes a sharp bend with its convex side out-
ward. To correct the disorders of the existing wall, it was
decided to build, under the cover of this structure, a geotex-
tile-reinforced structure having its facing set back about 20
cm from the uphill facing of the existing structure, designed
to take out thrust (Delmas et al., 1984).
 297

The structure is reinforced by nine layers of geotextile.

Because of its experimental character, three different geotex-
tiles were used: Propex 6066, Tri X 13221, and UCO 44615. We
give below only the results obtained on the first section,
which is the most fully instrumented.

The structure was built using one form per layer, supported
by the existing structure. Given the permanent character of
this structure and the state of knowledge when it was built,
the fill material chosen was a crushed limestone aggregate
having a continuous 0/3l.5-mm grading, placed at 95 % of the
normal Proctor optimum, or w = 5 % ; '( d = 20 kN/m 3 • The shear
strength was estimated to be : tf· = 45° ; c' = 0 kPa.

Because of the angular character of the fill material, a

compaction test was carried out. The main results, given in
table 4, show that there is little effect on the strength, and
that the modulus at failure is reduced by an average of 9 % in
the direction of stressing of the product.

Friction tests conducted using a large shear box ( 25 x

45 cm) yielded the following limits tan 0 = 0.78 for a soil/
geotextile relative displacement "up" = 2 cm.

The instrumentation implanted at the site included glued

strain gauges 10 mm long capable of measuring strains up to
10 %. The gauges were bonded to the geotextile on a rubber ce-
ment that ensured a suitable surface condition. A laboratory
calibration was carried out because we had little experience
with gauges of this type and, in particular, this way of bon-
ding them. This revealed, notably, that while in the short term
the measured strain values were reliable up to 2 %, in the
longer term the creep of the cement made adequ&te precision
impossible.

In addition, the deformations of the facing were measured

using inclinometer tubes set in PVC tubes placed on the outside
of the facing and attached by straps anchored in the structure.

Figure 5 shows the strains measured in the sheets when the

road was reopened. It can be seen that the measured strains do
not exhibit the distribution normally expected in reinforcing
structures, and in particular exhibit no maximum. Moreover,
the measurements made immediately after the removal of the
form from the layer corresponding to the sheet measured show
that, at this stage, from 70 to 95 % of the final strains have
been reached, with the balance appearing when the the next
layer is placed (table 5).
298

Table 4 - Langres : tensile properties of thegeotextiles

before and after compaction

Before compaction Loose after compaction

et - et £
-et -£ 0 0
- £
o (kN/rnl
0
(tl
-et (tl
£
(tl
0 0

t pp 77 ,3 18,3 0 - 10,0

z.

'3 . 00 __
~I
a strain after construction of layer
• strain at end of c~~struction

Figure 5 - National Highway 19 (Langres): strains measu-

red in geotextiles immediately after the cons-
truction of the pavement and reopening of the
road
 299

Table 5 - Langres : evolution of strain mobilization in

the layers of geotextile as the embankment was
built up

Point of measurement 1 2 3 4 5 6

Geotextile 2

E. (%) 1.52 1.10 1.36 1.56 1.28 1.44

~

Ef (%) 1.63 1.34 1.61 1.65 1.45 1.63

0.93 0.82 0.84 0.94 0.88 0.88
Ei/Ef

Geotextile 4

E. (%) 0.82 1.07 1.07 1.22 1.37 1.37

~

Ef (%) 1.05 1.40 1.51 1.54 1.67 1.59

0.78 0.76 0.70 0.79 0.82 0.86
EJE f

Geotextile 6

E. (%) 0.73 0.71 1.40 0.78 1.04 0.76

~

Ef (%) 1.07 1.03 1.60 1.04 1.34 0.94

0.80

.
0.69 0.68 0.87 0.75 0.77
Ei/Ef
,
E. strain of geotextile at the ehd of layer construction
~

Ef strain of geotextile at the end of wall construction

It would seem, in this particular case, that the placement

of the fill soil, and especially its compaction, account for a
large share of the final strains measured in the structure.

As it happened, the absolute necessity of ensuring the

verticality of the structure, to join up with the cantilever
footway, together with the problems encountered when the form
was removed from the first layer (because of slightly insuffi-
cient compaction), led the contractor to take special care in
the compaction of the fill, especially near the facing. In
addition to the desired density, this compaction resulted in a
lateral deformation of the structure towards the wall in the
layers from which the forms had already been removed. This
deformation resulted in tensioning of the geotextile, which
thereby confined the fill. This may explain the relatively
uniform strain along the sheets of reinforcement.

It will also be noted that the phenomenon seems to have

been enhanced by the fact he final strain valuees were reached
300

as soon as the next layer was placed. The situation might be

substantially different in the case of a very high embankment
(at least in the bottom of the structure), or if the top of the
structure carried a large surcharge.

From these few examples, it is manifest that the placement

method has a large impact on the strains measured in the rein-
forcing sheets. However, other factors may further modify the
strain distribution field in the reinforcements, such as
particular external stresses applied to the structure or
displacements of the facing.

To illustrate the former possibility, we may cite the

example of the Lixing structure (Delmas and Matichard, 1986).

2.4 - Lixing structure

To repair a shear slide that occurred shortly after the

excavation of a cutting for a new lane, a solution consisting
of the building of two super-imposed geotextile-reinforced
walls was proposed.

We shall deal here in particular with the upper structure,

the function of which is to retain the slide mass that could
not be excavated in the course of the work. To accomplish

this, the structure, which has a total height of 4.5 m and

is 5 m wide, is embedded 1.5 m in the sound soil (figure 6).
With respect to the slide, it acts as a stabilizing abutment.

However, the presence of horizontal sheets of geotextile

along which the shear strength is substantialtY less than in
the fill makes it possible for the rupture to propagate at the
geotextile. This to some extent complicates the design calcula-
tions for the structure by making it necessary to estimate the
forces induced in the structure by the slide mass and the
magnitude of their role in the internal behaviour of the
structure.

In the case of conventional abutments built with frictional

materials, the forces induced on the repair by the slide mass
may be estimated by a stability analysis along an imaginary
slip surface prolonging the existing surface inside the
abutment.

In the case of geotextile-reinforced abutments, this seems

to be the most realistic approach, since the structure acts on
the slide not as a rigid retaining structure but as an abutment
capable of taking out a large shear stress.
 301

initial profile
of cutting

3 .......
-t-t-
'~.
......,u.
3.'1'1'>

Figure 6 - Lixing : cross-section of proposed repair and

slip surface

The geotextile used is woven strips of UCO 44614 polypropy-

lene and the fill a sand (Vosgian sandstone) having the
following mean mechanical properties: c' = 0 kPa, 4-" = 35°
at 95 % of the normal Proctor optimum, or td = 17 kN/m 3 •

compaction tests on the geotextile revealed no significant

changes in the properties of the product.

The instrumentation of the structure consisted of glued

strain gauges to measure the strains of the geotextiles and an
inclinometer placed in the reinforced structure to measure its
displacements (figure 7).

The conclusion to drawn from these measurements is that

on the whole the strains in the geotextiles increased up to
about ten days after the construction of the abutment. The
inclinometers, on the other hand, indicated deferred strains:
practically no displacement at the end of construction, so
showing that the strains measured by the gauges were related
primarily to the internal stability of the structure; but, in
the year following construction, up to July 1986, an overall
displacement by 2.5 mm of the part located above the geotextile
extending the old slip surface. This finding confirms the
302

validity of designing the structure as an abutment, and also

reveals the gradual mobilization of friction along the sheet of
geotextile.

Although the strain gauges were no longer operational when

the displacements were measured on the inclinometer tube, it is
highly likely that these movements engendered in the sheets
tensions distributed differently from those observed in the
first stage. Finally, it will be noted that no significant
movement of this structure has been observed since July 1986.

Otherwise, the deformations found at the facing, if large

enough, may induce large modifications of the tensions in the
reinforcements. In what follows, we shall consider only
structures with flexible fari.ngs, the most likely to be
deformed.

+t z.s ~~ between 3/85 - 7/86

-- -- --I -- I

4.S'nl.. .

I
L
.. strain gauge

:'nclinometer • strain at 12/84

Figure 7 - Lixing : arrangement of measuring instruments

and valuesof displacements and strains measured
at the end of construction
 303

Combined measurements of the displacements of the facing

and of the reinforcing sheets have been made on only a few
structures. Among them we may mention the Allevard structure,
the Grenoble University Hospital Centre structure, both in-
strumented and monitored by the University of Grenoble, and a
structure on the A7 motorway at La Galaure, monitored by the
Lyon Laboratory. Although the facing measuring equipment
used on the first two structures is not without interest, we
shall consider here only the processing of measurements of the
last structure, in which the observed movements have been
found to affect the overall behaviour of the structure.

2.5 - Structure on A7 motorway

For the widening of the A7 motorway between Saint Rambert

and La Galaure, an approach combining geotextile reinforcement
with a prefabricated wall was chosen for the embankment
pertions. The structure consists of a soil mass reinforced by
six layers of woven polypropylene geotextile (UCO 44615)
together with a polypropylene-ethylene grid placed on the fa-
cing to retain the topsoil on the embankment slope, 50 0 from
the horizontal. This structure, which has a mean height of
2.7 m, is topped by a prefabricated wall 1.5 m high to allow
piping to pass (photo 4). The fill is a pea greavel having good
mechanical properties: c' = 0 kPa, If' = 41 0 , for a placement
density '( d = 22.5 kN/m 3 •

The properties of the geotextiles used are summarized in table 6

The instrumentation placed on the structure (Marchal,1987)

includes measurements of reinforcing sheet strains by inductive
sensors, vertical strain measurements by horizontal inclinome-
ters, measurements of the relative displacements of the
reinforcinggeotextile and the grid, and measurements of the
rotation of the underlying retaining wall. The inductive sen-
sors used here are attached to the geotextile by distribution
plates 30 cm wide attached to the geotextile 50 cm apart. The
long measurement baseline substantially decreases the uncer-
tainty arising from the fact that the geotextile is not flat,
but the main advantage of these sensors is good long-term
reliability.

Figure 8 shows the strains measured on the sheets and the

settlements of the sheets at the end of construction and six
months later. It will be noted that the lack of ties between
304

the geotextiles and the facing grid, together with the low
coefficient of friction of the two products, results in a
relative slippage of about 2 cm. The local decompaction at the
facing leads to local settlements, together with a small local
increase in the tensile forces near the end of the reinforcing
sheet and a slight rotation of the upper retaining structure.

Photo 4 - A7 motorway view of structure during construction

Table 6 - A7 motorway at La Galaure mechan~cal properties

of the geotextiles used

Tensile caracteristics
(NF 38014)
Ci (kN!ro) S- ('1;)

Geotextile
t PP 72.1 13.0

Geogrid
9.1 8.0
G PP PET
,
 305

6e~ o.15~

geogrid

1-
'f
.A (em..)

geotextile
3.5"..."...
settlement

strain

Cl~
0.6

0.2-
o
0.6

0.2.

0--0 end of construction (04-86)

.-.. (10-86)

Figure 8 A7 motor way arrangement of measuring

instruments and measurements made at the end
of construction and six months later
306

It may be concluded from this particular example that

proper compaction of the structure, in particular of the
facing, is the key to the proper behaviour of the structure,
and that the "reinforcing" and "facing" geotextiles must be
tied together to confine the soil and pretension the sheet.

The conclusion to be drawn from the few examples given here

is that while the tensions in the reinforcing sheets depend to
a large extent on the internal stability of the structure,
they are significantly influenced by the placement method and
by the nature and timing of the stresses to which the
structure is subjected. For this reason, the placement method
must be chosen not only to facilitate the building of the
structure, but also to ensure optimum tensioning of the sheets.
The method mentioned above may be used (Delmas et al.", 1986).
Since each layer is placed behind a form supported by a fixed
reference independent of the structure, it ensures a good final
geometry and effective compaction of the fill near the facing.
In addition, the absence of forms on the underlying layers
allows deformations of the structure by the compaction and
contributes to the pretensioning of the sheets. Along the same
lines, the process proposed by Mc Gown et ale (1987) promotes
tensioning of the sheets at the time of placement and serves to
reduce deferred deformations of the structure.

3 - THE USE OF CLAYEY SOILS IN GEOTEXTILE-REINFORCED STRUCTURES

In so far as it is not necessary to use dilatant soils in

geotextile-reinforced structures, soils having a large frac-
tion of fine materials can be used. If the soil used meets the
usual specifications for fills, its use in a reinforced
structure does not, a priori, pose any special'problems if its
mechanical properties are properly taken into account in the
design and if the water content of the soil at placement is not
likely to result in pore overpressures during subsequent
loading.

The example of the A15 motorway, mentioned above, is a case

in point. In spite of placement in a moist condition (Sr = 94
to 98 %), which resulted in a rubber-cushion effect during
compaction, no disorders were found at the time of placement.
This may be explained by the use of a draining geotextile -
here, a needle-bonded nonwoven - that served to prevent the
creation of large pore overpressures and facilitated the conso-
lidation of the fill material. The traces of deposited calcite
on the samples taken in 1986 seem to confirm the draining
action of the geotextile.
 307

The example of the experimental embankment at Rouen provides

some additional information about the actual behaviour of
reinforced fine soils having a high water content (Blivet et
al., 1986). This experimental structure, 5.6 m high, was built
with a silt having a water content of wnpo + 5 % and was
reinforced with various types of geotextiles. Here we shall
consider only two of them, on which pore pressure measurements
were made : an Enka woven polyester and a needle-bonded nonwo-
ven polyester in conjuction with a Rhone Poulenc grid.

o- - - 0 G NTA PES pore pressure

0-'- Q t PES
1 50
(k Pd.)

2.0

----t--'O

• Jo

I
I
:::0-1

Figure 9 - Rouen experimental structure : arrangement of

measuring instruments and measurement values
120 days after construction

Because of the type of soil used as fill, its poor mechani-

cal properties, and the type of forms used, the placement
strains ' were large.

Considering mainly the measurements made from the end of

construction, figure 9 shows the difference between the
behaviour of the products. The pressure sensor inside the
308

embankment, outside the structures, reveals placement overpres-

sures of as much as 50 kPa.

On the woven sheet, the pore pressures are positive at the

back of the structure, then disappear and finally become
negative near the facing.

On the composite geotextile, on the other hand, the pres-

sures are negative over the whole length of the reinforcement.

This difference in local behaviour can lead to large

changes in overall stability on a nearby test section
reinforced with a woven polyester with its surface treated to
be non-wetting, the soil mass turned over because of anchorage
failure. An after-the-fact calculation revealed an effective
angle of friction of 5°, as against a soil-geotextile angle of
friction of 21° in a drained condition.

We may conclude from these two examples that, while the use
of soils having a large percentage of fines poses no special
problems, it is necessary to make sure that the placement and
compaction are not likely to engender pore pressures, or that
the geotextile can dissipate them so as to ensure optimum local
soil-geotextile friction.

4 - INFLUENCE OF THE LONG-TERM BEHAVIOUR OF THE GEOTEXTILES ON

THE REINFORCED STRUCTURES

Without attempting an exhaustive analysis of the problems

of the long-term behaviour of the geotextiles, one may attempt
to learn something from a few examples.

With respect to creep behaviour, it will be noted that most

authors agree that creep in geotextiles depends to a large
extent on the load factor. In particular, some recognize that,
below 10%, creep in polymers at ambient temperature is
negligible. For guidance, table 7 gives the load factors of the
geotextiles at the sites mentioned earlier.It can be seen that
they are less than 10 % in all cases. Without attempting to
predict the future, it would therefore seem that in these few
structures the risk of creep is very small.

Moreover,the A15 motor way structure provides information

about the ageing of geotextiles. This structure, stressed for
eight months in 1971, has since been protected by a bank of
fill. The sampling done in 1986, or 15 years after placement,
provided an opportunity to measure the mechanical properties
 309

of the product. The loss of strength found, 21 %, agrees with

the loss of strength routinely observed during placement (up to
30 % with angular materials). And the gain in stiffness found
can be explained by the presence of calcite deposits in the
geotextile. Chemical analyses of the fibres revealed no change
of the crystalline structure (percentage, orientation),
confirming the absence of chemical deterioration of this type
of product in a soil of pH 10.

Table 7 - Maximum work factor measured on the geotextiles

at a few instrumented sites

Location Date Geotextile Soil a/af E

I

A 15 1971 NTA PES Clay

0,06 1,8 %
(Ph 10)

Langres 1983 t PP Limestone 0,08 1,6 %

Allevard 1983 t PP Gravel clay 0,9 2 %

Lixing 1984 t PP Sand clay 0,12 2,7 %

A 7 1986 t PP Gravel clay 0,05 0,6 %

This finding confirms the results of earlier studies based

on samples taken from actual sites (Sotton et al.! 1983).

5 - CONCLUSION

The conclusion to be drawn from these few instrumented

examples is that the tensile forces, and in particular their
distribution in the structure, depend to a large extent on the
conditions of placement, but also on the way the structure is
stressed.

This must be taken into account in designing these

structures, and for this reason preference should be given to
design methods in which the actual conditions of placement can
be simulated and taken into account. In this connection, we may
note the interesting approach made possible by the "displace-
ments" method (Gourc et al., 1986).
310

In addition, an attempt should be made to optimize the

mobilization of the forces by chosing a type of form that
allows pretensioning of the geotextile and holds deferred
deformations of the structure to a minimum.

REFERENCES

1) BLIVET J.C., JOUVE P., MAILLOT R. (1986) Numerical modeli-

zation of earth reinforcement by geotextile : hydraulic
function. C.R. IIIe Congo Int. Geotextiles, Vienne, avril,
IV, pp 1061-1066.

2) DELMAS P., FAVRE J.M.,MATICHARD Y., LEHMANN M., PRUDON R.,

REBUT P. (1984) Renforcement par geotextile d'un mur de
soutenement sur la RN 19 a Langres. Revue Generale des
Routes et Aerodromes, 609, juin, pp 61-66.

3) DELMAS P., MATICHARD Y.(1986) Landslides confortation with

geotextile reinforced earth work. C.R. IIIe Congo Int.
Geotextiles, Vienne, avril, IV, pp 1091-1096.

4) DELMAS P., PUIG J., SCHAEFFNER M. (1986) Mise en oeuvre

et parement des massifs de soutenement renforces par des
nappes.

5) GOURC J.P.,RATEL A., DELMAS P.,(1986) Design of fabric

retaining walls: the "displacement method". C.R. IIIe
Congo Int. Geotextiles, Vienne, avril, IV, pp 1067-1072.

6) MARCHAL J.(1986) Autoroute A7 : Elargissement entre St

Rambert d'Albon et la Galaure. Compte rendu de mesures.
Rapport Labor. P. et Ch., 30 pp.

7) Mc GOWN A., ANDRAWES K.Z., MURRAY R.T. (1987) The influ-

ence of lateral boundary yielding on the stresses exerted
by back-fills. C.R. colI. Interactions Sols-Structures,
ENPC, mai, pp 585-592.

8} ~ERRIER H., LOZACH D. (1986) Essai de poin90nnement sur

geotextile - Influence des sollicitations de compactage.
Compte rendu de mesures. Rapp. Labor. P.et Ch., 45 pp.

9) PUIG J., BLIVET J.C. (1973) Remblai a

talus vertical arme
avec un textile synthetique. Bull. de Liaison P. et Ch.,
64, mars-avril, pp 13-18.
 311

10) SCHLOSSER F., GUILLOUX A. (1981) Le frottement dans le

renforcement des sols. Revue fran9. de Geotech., 16, pp
65-77.

11) SOTTON M., LECLERCQ B., PAUTE J.L., FAYOUX D. (1982)

Quelques elements de reponse au probleme de durabilite
des geotextiles. C.R. lIe Cong.lnt. Geotextiles, Vol. 2,
Las Vegas, aout, pp 553-558.
GEOTEXTILE REINFORCED RETAINING WALLS: DISCUSSION OF INSTRUMENTED LARGE
SCALE TEST WITH RESPECT TO THE VERIFICATION OF DESIGN CONCEPTS

B. GRAF, J.A. STUDER

GSS GLAUSER STUDER STUESSI, CONSULTING ENGINEERS INC., ZURICH, SWITZERLAND

1. INTRODUCTION

To conserve the landscape in the construction of new arterial roads it is

becoming more and more necessary to employ deep cuts with a long service
life and high reliability using steep artificial slopes for stabilization.
Such slope stabilization can be built with relatively low expenditure, if
the excavated material is used as fill material and the fill material is
strengthened by geotextile reinforcement. High standards are required for
the safety and durability of such permanent slopes. Besides the fact of
relatively low costs for geotextile reinforced retaining walls (compared
to the expenses for conventional retaining walls), the total construction
expenditure can be high (long and deep cuts). The above mentioned require-
ments are arguments for the need of a consistent, verified design concept
for geotextile reinforced slopes. To the best of our knowledge, such a
concept does yet not exist.

On the other hand, geotextile reinforced structures have been built

successfully for a long time. However, with respect to safety and relia-
bility, this construction method has been mostly used and accepted for
less important structures (for example low and/or temporary embankments).

Against the application of geotextile reinforcement for high and steep

slopes, however, with a service life of 50-100 years, the following argu-
ments are used:

- Insufficient permanence of geotextiles against UV-radiation,

- Susceptible in the case of acts of sabotage,

- Insufficient knowledge about ageing and related changes of material

strength for geotextiles,

- Aesthetics,

- Maintenance.

Mostly it is suggested, to compensate for these almost incontrollable in-

fluences on the stability and durability of such structures, by appropri-
ate reductions of the material strengths in the static analysis, appropri-
ate construction methods (lining) and with the help of a long term monito-
ring of such structures. However, appropriate reduction of material
strengths presumes realistic results for the forces and stresses from the
static ananlysis.

313

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 313-337.
© 1988 by Kluwer Academic Publishers.
314

The limit analysis design of geotextile reinforced retaining walls is

usually similar to that of conventional retaining walls (cf.(8)), whereas
the application of such conventional analysis on reinforced walls has not
yet been verified.

Because of the high costs of large scale tests, only a few fully instru-
mented and evaluated cases have been reported. For none of the published
experiments limit states could be reached (cf.(13),(15)). The design,
however, is mainly based on limit analysis.

From the foregoing it is clear that there is a need for additional theore-
tical and experimental investigations of the behaviour of geotextile rein-
forced retaining walls.

The aim of this paper is the discussion of available limit analysis design
concepts and their experimental verification. In section 2 for several
reasons mainly the design concept proposed by Gudehus and Schwing (8) will
be presented. The reasons are: Concept is based on model test results, the
most important aspects of stability are incorporated, partial factors of
safety have been calculated with a probabilistic method by these authors.
Section 3 deals with the experimental verification of the presented design
concept, whereas a procedure for test evaluation is discussed, followed by
the application of this procedure to the evaluation of published data from
large scale tests. Finally, in section 3 the proposals for future theore-
tical and experimental investigations are summarized. Section 4 deals with
case studies.

2. DESIGN OF REINFORCED ARTIFICIAL SLOPES

Design steps are discussed with respect to the standard reinforced wall
shown in figure 2.1: cohesionless soil, no surcharge loads. This section
contains the theoretical background needed for evaluation of field data.

FIGURE 2.1. Standard geotextile reinforced retaining wall.

Today's design of reinforced artificial slopes is mostly based on limit-

analysis models describing the failure of the structure. For displacement
estimates empirical formulas are used. In the following the limit analysis
desig~, determination of displacements and the statement of partial
factors of safety are discussed.

2.1. Limit states

For convenience, limit states will be discussed separately for modes of so

called internal-and external failure. External failure is characterized by
 315

failure along distinct slip surfaces, whereas all the other failure modes
(for example zone failure) are called internal.

2.1.1. External failure. For external failure modes usually rigid-body

failure mechanisms are assumed. Figure 2.2 shows some of these mechanisms.
Figure 2.2a holds for failure along a cylindrical slip surface. Various
design charts are based on this mechanism. These charts can be used for
the determination of the design length and tensile strength of reinforce-
ment (e.g. (5)).

a) b)

c) d)

FIGURE 2.2. Failure for internal stability: a) Rotation, b) Translation,

c) Translation, 2 Rigid bodies, d) Sliding.

English codes assume the Coulomb wedge from figure 2.2b. This kind of
failure can be expected for embankments with local surcharge loads (10).

The so-called Dutch/Enka-Method (4) assumes a failure mode of 2 rigid-

bodies under pure translation (figure 2.2c). This failure mechanism was
originally observed in model and field tests concerning soil-nailing (7).
Figure 2.2d shows the assumed failure mechanism for sliding. This kind of
failure can be expected for insufficient design length of the reinforce-
ment.

To the best of our knowledge, none of the discussed failure mechanism has
been verified yet by experiments for geosynthetic reinforced retaining
walls. For reasons of safety (in the sense of the so-called collapse
theorems of the theory of plasticity), however, for kinematic solutions,
the actual failure mechanism should be well approximated by the assumed
one.
Gudehus und Schwing (8) published a paper with model test results for the
failure mechanism. In these tests the stability of standard walls
(cf. figure 2.1) was investigated as a plane strain problem giving regard
to the geometrical and mechanical rules of similarity. The height of these
walls was increased until the collapse of the structure occured. The
observed failure mechanism (cf. figure 2.3) is different from the usually
assumed ones shown in figure 2.2. For walls with no surcharge loads,
cohesionless soil and sufficient internal stability, therefore, a failure
mechanism of 2 rigidbodies under pure translation, can be assumed. More
316

specific features of the observed mechanism will be discussed later.

h ~I----:

.1
FIGURE 2.3. Observed failure mechanism (Gudehus and Schwing, (8)).

Rotational mechanisms or the Coulomb-wedge can be expected for restricted

surcharge loads. For sail-nailing under these conditions (restricted sur-
charge loads) a mechanism of 2 rigid-bodies under rotation was observed
(8). A large scale test from (15) supports the Coulomb-wedge assumption.
Experimental verification is planned for geotextile reinforced retaining
walls (8).

From model test observations and numerical studies the following knowledge
has been obtained for the failure mechanism shown in figure 2.3 (cf.(8)):

- The slip surface 24 coincides with the ends of the reinforcement. This
observation simplifies the formulation and evaluation of the limit state
equation.

- For the angle ~? (cf. figure 2.3) the Coulomb solution can be assumed
(1J1 = 45° + tp/ti: ; f: friction angle) . This result was derived by nume-
ricaf studies and was experimentally verified.

- The failure mechanism from figure 2.3 contains the following failure
modes as particular cases: overturning, sliding and bearing (vertical
translation). For example, it can be shown with the help of the limit
state equation, that for small Ah.lh. ~ '11.;'_0' • holds, i.e. in
this case sliding is the relevant failure mode for the structure.

- Because of local base failure, the vertical stresses at the base of the
wall are essentially smaller than the overburden-pressure (cf.(8), (3)).
As a consequence of this experimental finding, Gudehus and Schwing (8)
neglect in stability calculations the pull-out resistance of the rein-
forcement near the toe of the wall (up to 10% wall-height). In all other
layers the pull-out resistance is:
(2.1)
where

overburden-pressure for reinforcement layer i,

length of reinforcement layer i behind slip surface,
friction coefficient (soil/geosynthetic).
 317

With the help of the abovementioned findings, the limit state equation can
be easily formulated and evaluated. In this equation the slip-surface in-
clination r~ (cf. figure 2.3) occurs as the only unknown quantity. This
inclination can be derived with the assumption, that the sum of the ten-
sile forces should be maximum along the slip surface ~. This assumption
has been experimentally verified (8). With the help of the abovementioned
limit state equation, a design chart was made available by (8). This
design chart is further discussed in section 2.3.

Other stability investigations, based on one of the failure modes shown in

figure 2.2, will not be further discussed here, because they have not been
experimentally verified. As already mentioned, the investigations of (8)
are restricted to the standard structure from figure 2.1. The incorpora-
tion of surcharge loads is planned in future work by the last mentioned
authors.

2.1.2.Internal stability. The following, so called internal failure

modes will be discussed:

- Failure of lining due to earth pressure,

- Failure of reinforcement.

For design purposes the earth pressure and the maximum tensile force
(related to internal stability) for the reinforcement must be known.

Field data for the earth pressure derived by (10) imply, that the use of
active earth pressure loads leads to conservative design (cf. figure 2.4).
For the upper part of the wall the measured pressures are only a little
greater than the theoretical active pressures; for the lower part of the
wall, however, the measured pressures are up to 50% lower than the active

r
ones.

II
h

IL ax ax
1
\
\
\

ax
0) b) c)

FIGURE 2.4. Earth-pressure distribution: a) h = 2 m, b) h 4 m,

c) h 7 m. Dashed line: Active earth pressure (cf. (8)).
John (10) calculated the earth pressures, shown in figure 2.4, from the
measured connector-forces (the reinforcement is fixed to the lining with
the help ;f connectors). Such force-measurements are much more reliable
than measurements with the help of earth pressure cells.

With increasing distance from the lining, the forces in the reinforcement
increase (cf. (16) and figure 2.5), i.e. the maximum tensile forces are
not equal to the earth pressure forces.
318

5.0

4.0
~
3. °
.""
~

W2
..::0.:...O"-~=-~-::-:.::.l-:"""--;;. ____ _
o 1 2 3 -4
Distance from slopf" surface (m)

FIGURE 2.5. Measured strain s for the reinforcement (Yamanouchi et al.,

(16)) •

Firstly, the design earth pressure will be discussed. Most standards re-
commend a linear earth pressure distribution (cf. (10)). Gudehus and
Schwing (8) assume a linear distribution down to 2/3 of the height of the
wall and afterwards a constant value down to the toe (cf. figure 2.6). The
inclination of the linear part of the distribution should be assumed like
in the active case. The distribution from figure 2.6 was chosen as an
approximation of the field data derived by (10). A more sophisticated
approximation of the field data can be derived with the help of the
so-called silo theory (8). By the assumption of an linear distribution,
the pressures are clearly overestimated (cf. figure 2.4). The proposed
distribution of (8) is more realistic, easy to use and,therefore, will be
further discussed in the context of the evaluation of field data (cf.
chapter 3).

A'V<'«'<

h
fAh

FIGURE 2.6. Design earth-pressure distribution (Gudehus and Schwing, (8)).

As mentioned before, the tensile forces in the reinforcement can increase

essentially with growing distance from the wall face (cf. figure 2.5).
There is as yet no theoretical prediction discribing this increase. Mea-
surements from (15) indicate that the maximum tensile force can be up to 4
times higher than the resultand earth pressure force acting on the wall
face •.Gudehus and Schwing (8) assume the following correlation between the
earth pressure and the design tensile force:
mq)< 1:" O.~?l' f-t. Ah., K",J,.; (2.2)
Kal.. '" +-g.'2 <. '-t 5° - ¢/2.) i (2.3)
i.e. the maximum tensile force is assumed to be correlated to the earth
pressure for the depth of O.9h (cf. figure 2.2). This correlation has not
 319

yet been verified by experiments.

2.1.3.Factors of safety. From the literature very little direct informa-

tion can be gained about the factors of safety to apply for the design of
geosynthetic reinforced walls.

For the design of reinforced retaining walls, the partial factors of safe-
ty for the material properties of soil, the friction coefficient between
soil and geotextile, the material properties of the reinforcement and the
surcharge loads are needed.

For the case of external stability, (8) calculated the partial factors of
safety for the standard wall shown in figure 2.1. They assumed the fric-
tion coefficient and angle of friction, occurring in the limit equilibrium
equation, to be statistically independent, random quantities (cf. (8)), for
the assumed probability distributions) and derived the partial safety fac-
tors with the help of the second-moment reliability code. With respect to
the mean values of friction angle and friction coefficient, the following
results were derived:

For the friction angle (of the soil) the partial safety factors
are in the range of 1.1 - 1.2, for the friction coefficient in
the range of 2-4. These ranges for the safety factors correspond
to the range of 30°-40° for the mean values of friction angle of
the soil.

Under the assumption that the tensile strength of the reinforcement

exceeds the pull-out resistance, the tensile strength does not occur as a
variable in the limit equilibrium equation. Gudehus and Schwing (8),
therefore, assume a partial factor of safety for the calculated tensile
force, within the range 3-4. The design tensile force has to be compared
with the tensile strength of the reinforcement~ In addition, to avoid
creep of the reinforcement, the corresponding strain should be less than
5% (cf. also (17)).

The partial factors of safety proposed by (8) have to be verified by back-

calculation of the reliability of well performing reinforced retaining
walls. Besides this, surcharge loads should also be incorporated and the
assumption of statistical data (distribution and corresponding parameters)
should be verified with the help of experiments.

Besides the applied second-moment reliability code, other probabilistic

calculations are possible (cf. (17)). The second-moment code, however,
seems to be the most suitable method:

- Reliability of different types of constructional methods is comparable.

- Derivation of the widely accepted and easy to use partial factors of

safety, on the basis of a consistent probability concept.

The above discussion holds for the partial factors of safety corresponding
to external stability., For the case of internal stability and deforma-
tions, to the best of our knowledge, similar investigations do not exist.

By limiting the strains in the reinforcement, caused by the calculated

tensile forces, to small values ( 5%), creep-deformations can successfully
320

be prevented (cf. (17) and (16». This holds for cohesion less fill
materials and high strenght reinforcement, characterized by an initially
steep inclination of the stress-strain curve. The creep problem, there-
fore, is mainly relevant for cohesive fill materials and/or continuous
filement needle punched geotextile as reinforcement, which are characte-
rized by a relatively low initial strength.

2.2 Displacements

2.2.1 Empirical estimates and possibilities of theoretical predictions.

The total displacements of a geosynthetic reinforced wall can be separated
as follows:

- Settlement of the contact area,

- Deformations of the wall during construction,
- Deformations due to surcharge loads,
- Deformations caused by creep under gravitiy and surcharge loads.

The settlements of the contact area can be calculated conventionally and,

therefore, will not be further discussed.

By experience it is well known that the largest deformations occur imme-

diately (if creep is prevented). The amount of these deformations is
largely influenced by the method of construction (cf. (8), (3D. The dis-
placements of the head of a wall can reach up to 1-2 % of the height of
the wall (8). A theoretical prediction of these displacements can hardly
be done, also a prediction is not of great interest, because displacements
of these amounts can be accepted for this construction method.

A good prediction of displacements caused by surcharge loads is still not

possible. The application of numerical methods (Finite Elements) in the
near future - with the goal of quantitatively good predictions - is pre-
vented for several reasons (experience with constitutive equations, nume-
rics for nonlinear problems). The prediction of such displacements, there-
fore,also must be based mostly on experience.

By restriction of the expected strains of the reinforcem~nt (for the de-

sign tensile force) on small values « 5 %), creep deformations can
successfully be prevented (cf. (16». This holds for cohesionless fill
materials and strong reinforcements, characterized by an initially steep
inclination of the stress-strain curve. The creep problem, therefore, is
mainly relevant for cohesive fill materials and/or continuous filament
needle punched geotextiles as reinforcement, which are characterized by a
relatively low initial strength.

Displacements are mostly predicted with the help of the above mentioned
empirical estimates. Yamanouchi et. al. (16) published the results of a
FE-analysis predicting the displacements of a geotextile reinforced re-
taining wall under gravity and surcharge loads. For details of the mo-
delling the reader should refer to Yamanouchi et. al. (16). For this
retaining wall, the measured displacements were underestimated by a factor
of 0.5 by the FE-analysis. Creep was not investigated. Qualitatively, how-
ever, some of the derived results show good agreement with the measure-
ments:
 321

Distribution of tensile forces in a reinforcement layer,

- Distribution of tensile forces with respect to the height of the wall

(cf. figure 2.5).

2.3. The design concept

In this chapter a short overview of the design concept, proposed by (8),

will be given. This concept was derived for the standard retaining wall in
figure 2.1: cohesion1ess soil, no concentrated surcharge loads. This con-
cept will be used later on for the interpretation of field data.

The case of external stability will be dicussed first. In model tests the
failure mechanism, shown in figure 2.3, was observed. With the help of the
corresponding (to the failure mechanism) limit equilibrium equation, the
design length 1 of the reinforcement layers and the tensile force Z can be
calculated for the given geometry and soil data of a retaining wall. The
design quantities 1 and Z (maximum of Z according to external or internal
stability) are plotted in figure 2.7 versus the friction angle ~ , with~~/~
as contour lines. The design chart (figure 2.7) holds for friction coeffi-
cients (soil/geotextile):

(2.4)

i.e. the friction coefficient is already reduced by a partial safety fac-

tor (cf. chapter 2.1.3). Only the friction angle of the soil has to be re-
duced by a partial safety factor as mentioned in chapter 2.1.3, for the
application of the design chart.

~ Ah _
50 ~ ::::::----.- 11-
-----------
i, -..:::::: ~ 0,20
0,15
0,05/0,10
I
o 1,0 1,5
25 301 35
Ijl

0,01, 1---------_ .... -L

--
0,05

---
i t-----
I--" _
__ 0,10
0,15
0,20
- -::::::::::::::::---
0,08

::--.- I--
0,12
max Z
lh2
FIGURE 2.7. Design of a stande.rd geotextile reinforced retaining wall
(Gudehus and Schwing, (8»: Design chart (cf. figure 2.8 for situtation)
322

h
t~h 4l,Y

FIGURE 2.8. Si tuati9n corresponding to design chart

For reasons of internal stability, also the tensile force from equ. (2.2)
has to be calculated. The design, finally, has to be based on the largest
tensile force to the cases of internal or external stability. In figure
2.7 . themaximumtensile forces are plotted versus the friction angle of the
soil. This tensile force should be multiplied by a safety factor of 3-4
and these be compared to the tensile strength.

To avoid creep the strains caused by the tensile force should be at least
less than 5%. The appropriate value for this strain limit can be gained
from experiments.

According to experience, the displacements of the top of the wall can be

expected to be in the range of 1%-2% of the height of the wall.

3 .EVALUATION OF INSTRUMENTED DATA

3.1.Principles of evaluation

Most of the input and output quantities of the design concept (for example
tensile forces corresponding to limit states) cannot be measured directly
in full scale tests. In addition, in most large scale tests the limit
states which are assumed for design are not reached (cf. (13), (15)). In
the following, therefore, firstly the possibilities of verification of the
design concept are discussed.
3.1.1.Limit states

First of all, the case of external stability will be discussed. For veri-
fication of the design procedure, the following instrumented data and ob-
servations are needed:

- Geometry (height of retaining wall, length of reinforcement), loads and

material properties corresponding to limit states,

- Position and inclination of slip surfaces,

- Tensile forces for the reinforcement layers intersected by slip sur-

faces.

The geometry and loads can be easily determined. As mentioned before, how-
ever, l·imi t states can hardly be reached in large scale tests: Because of
large displacements, the loading equipment and instrumentation fail before
the limit state is reached (15). Large scale tests, therefore, are not
very suitable for the determination of the critical geometry and load.
 323

More suitable are model tests, which can be used for the systematic inve-
stigation of the behaviour of reinforced retaining walls in limit states
(cf. (8)).

Now the determination of the failure mechanism will be discussed. Figure

3.1 shows the measured geotextile strains from a large scale test, presen-
ted as contour lines (cf. (15)). In this test, the ultimate surface load
could not be reached. Therefore, the strains from figure 3.1 correspond to
a pre-failure state. The expected failure plane ist also shown as a
straight dashed line in figure 3.1. This line was drawn under the assump-
tion, that a failure plane would intersect the reinforcement layers at the
points of maximum strains. Even if the limit state was not reached in a
large scale test, the strain measurements, therefore, can be used for de-
riving a first estimate of the position of potential slip surfaces. More
conclusive results could be expected from data of slope indicator tests if
there are any. Because of the large expenditure, however, large-scale test
are not suitable for investigations of failure mechanismes. Model test are
the better "tool" for such investigations, whereas large scale tests
should be used for verification of model test results (adherence of simi-
larity rules).

p=500kN/m

E
...
.n.

552 921 S11 dot"'" 1400

1 + - - - - - - 3.25m

FIGURE 3.1. Contour lines for the geotextile strains (Wichter et al.,
(15)). Dashed line: expected slip surface.
324

Because of the large displacements, even for pre-failure states, it can be

expected, that the tensile forces in the reinforcement can hardly be mea-
sured for ultimate states. The tensile forces resulting from calculations,
therefore, can only be verified indirectely.

Large scale tests, therefore, are not well suited to the verification of
the modelling of external stability. Large scale tests should be performed
only in combination with model tests.

The design for internal stability depends on the following quantities:

Earth pressure distribution,

Maximum tensile force to be expected.

The earth pressure distribution, which is needed for the design of the
lining and connectors (connection between lining and reinforcement)', will
be discussed first. With the help of force measurements in the connectors,
the earth pressure distribution can be determined with high accuracy
(cf. (10); (8)). Predicted earth pressure distributions, therefore, can
easily be verified in large scale tests.

More difficult is the measurement of tensile forces in the reinforcement.

Usually the strains are measured. Because of the complicated material be-
haviour of geotextiles (creep, temperature-effects, strain-rate dependen-
cy, influence of normal stresses on the material behaviour of needle-
punched geotextiles) the back-calculation of forces is uncertain. Direct
force measurements would lead to more reliable results. For the design of
the reinforcement only the maximum tensile force is needed. The presented
design concept assumes a correlation between this maximum tensile force
and a resulting earth pressure force (cf. equ. (2.2)). This correlation
has to be verified by experimental data. This data can be gained from
large scale tests.

The pull-out resistance of reinforcement layers, incorporated in the limit

state equation for external stability, depends strongly on the vertical
stress assumed to act at a certain depth (cf. equ. (2.2). There are very
little experimental data available for the distribution of vertical stres-
ses with depth. Experimental data indicate that the vertieal stress is
considerably below the overburden pressure for points lying near the face
of the wall. The assumption of a hydrostatic distribution for the vertical
stresses with depth, therefore, would lead to an overestimation of the
pull-out resistance. For this point to become more clear, measurements of
the vertical stresses from large scale tests are needed.

3.1.2 Displacements. Up to now, the displacements cannot be predicted

with the help of easy to use analytical or numerical methods. The dis-
placements caused by gravity loads can be estimated with the help of empi-
rical formulas. Creep is usually prevented by definition of a limit for
the geotextile-strains to be expected under the calculated tensile forces.
These formulas and limits will be verified later in connection with the
evaluation of field data.

3.1.3 Safety requirements. Most of the papers on the subject of rein-

forced retaining walls deal with investigations of the statical aspects of
external stability. There are only few papers, however, dealing with
 325

safety requirements and proposals for the use of partial factors of

safety.

One of the open questions concerns the basic statistical data needed for
the application of the second moment reliability code. Thus, very little
is known about the probability distribution of the pull-out resistance.
Also, back-calculations of the reliability of existing, well performing
retaining walls have not yet been published. The question concerning par-
tial safety factors, therefore, is closely related to experimental inve-
stigations. Future work on the subject of geotextile reinforced retaining
walls, therefore, should also concentrate on the following topics:

- Determination of basic statistical data (type of probability

distributions and their parameters),

- Documentation of well performing structures and back-calculation of'

reliability (cf.(2)).

3.2 Experimental results and predictions

From large scale tests or existing walls mostly only data for displace-
ments are available. Only few data are available for verification of the
force and pressure quantities resulting from the design concept for the
case of internal and external stability (cf. table 3.1).

Authors measured
displace- tensile soil observat'ed remark
ments forces pressures failure
mechanism

Yamanouchi et al. x x - - -
(16)

Caroll and - x x - -
Richardson (3)

John (10) - - x - only

lateral
pressures

Wichter et al. (15) x x - x cohesive

fill ma-
terial

TABLE 3.1. List of published large scale tests.

For none of the large scale tests listed in table 3.1, data corresponding
to the limit state are available. For some of these tests limit states
should not be reached, the test reported by (15) had to be terminated for
technical reasons before the limit state was reached. Consequently, the
design concept concerning internal and external stability cannot be veri-
fied yet with the help of measurements from large scale tests. On the
326

other hand, however, some of the available data indicate that the ultimate
surface loads are strongly underestimated (cf. (15); (13». This problem
will be discussed in the next section.

The data from the first 3 tests, listed in table 3.1, correspond to ser-
vice states. These data can be used for verification of the predictions
for the earth pressure and maximum tensile force corresponding to internal
stability (cf. section 2.1.2).

Finally, the predicted and measured displacements are compared.

3.2.1 Limit state. Several results from large scale tests indicate that
the ultimate surface loads are considerably larger than the predicted ones
(15); (13). The discussion of this contradiction between experiment and
prediction will be based on the experimental and theoretical findings of
Wichter et. al.(15).

Figure 3.2 shows the structural details, the position of the surface load,
and the expected failure surface (cf. chapter 3.1.1) for the retaining
wall. T~e test was §arried out with cohesive fill material: = 21.5°, c = ¢
40 kN/m ,t=20 kN/m . A Stabi1enka 200-type reinforcement was used.

KEY PLAN G; 90 kN/m

C , 200 kN/m

5, 014.37.0.36 P IkN/ml
S:1 ,42.86 .0.46 P
53 ' 85.06 • 0.59 P
R
54 ' 141.16.0.71 P
max 5 0 263.5 .2.12 P

lcm ='OOkN/m
200 IJXJ

FIGURE 3.2. Prediction of ultimate load and tensile forces: Situation,

slip surface, polygon of forces.

At the end of the test, a surface load of P = 500 kN/m was applied, but
the ultimate state was not reached under this load.

Wichter et. al.(15) used the failure mechanism from figure 3.2 to predict
the ultimate surface load. For the surface load a load distribution was
assum~d. Consequently, the total pUll-out resistance (cf. Sl -S4 in figure
3.2) can be calculated as the sum of the following terms: A constant term
due to gravity loads and a term which depends on the surface load. The
formulas for the pull-out resistances are given in figure 3.2. Further-
more, in figure 3.2 G stands for the self-weight of the failure wedge, C
for the cohesion-force on the failure plane and R for the resulting
 327

friction force on the failure plane. For the calculation of the ultimate
surface load, the resulting pull-out resistance (max S = 283.5 + 2.12 P;
cf. figure 3.2) was assumed. Formulation of the equilibrium equations and
solving the equations for the unknown surface load P leads to the follo-
wing result: Only for P< 0, i.e. tensile forces as surface loads, equi-
librium can be reached. Wichter et al. (15) gives the following interpre-
tation for this result: The reinforcement layers (for the assumed pUll-out
resistance) at the limit state cannot be pulled out all at once, i.e.
there must be a so-called mechanism of intrinsic resistance.

On the other hand, each of the tensile forces Sl to S cannot exceed the
tensile resistance of the geotextile (200 kN/m). Base~ on the information
in figure 3.2 the pull-out resistance is greatest at position S4. Wichter
et al.(15) conclude from this that the tensile strength is first reached
in position S4 and then assume S4 = constant = 200 kN/m. With this assump-
tion a value of ultimate surface load of P = 83 kN/m can be calculated;
i.e. it is very much smaller than the value observed experimentally
(P = 500 kN/m). In addition, in the test the limit state was not reached.
Thus there is a big contradiction between predicted and observed results.

This contradiction, however, is caused by a statical misinterpretation.

Assuming max S (cf. figure 3.2) for the resultant pUll-out resistance, the
polygon of forces can be closed, if an additional force, the so-called
"cause of failure-force" F f (cf. (6)), is introduced (cf. figure 3.3). The
polygon of forces, now, can be interpreted as follows: Failure occurs for
a reduction of the pull-out resistance by the value of the force F f , or,
under the given gravity and surface loads the structure is stable, and an
appropriately defined factor of safety, for the "pull-out case", will have
a value greater than unity. This does not contradict the experimental
findings.

max S G
c

FIGURE 3.3. polygon of forces, failure caused by decrease of pull-out

resistance.

Also, failure caused by exceeding the tensile strength of the reinforce-

ment has to be examined, which will lead to a different value for the
factor of safety.

For design purposes, the following problem has to be solved:

328

1. Determine, for a given surface load P, the resultand tensile force of

the reinforcement, which corresponds to the limit state.

2. Choose the material and geometrical data of the reinforcement in such a

way that neither the tensile strength nor the pull-out resistance is
exceeded by the predicted tensile force.

The solution of problem 1 leads to the following equation for the tensile
force:

(3.1)

l : shear-band inclination.

Using equ. (3.1) with the input-data: C = 200 kN/m, ~ = 55°, ¢ = 22.5°,
P = 500 kN/m (largest applied load in test), G = 90 kN/m (self-weight of
the failure wedge), SR = 156 kN/m can be calculated for the resultand ten-
sile force. This tenslle force is even smaller than the tensile strength
(200 kN/m) of a single reinforcement layer and very much smaller than the
maximum available pUll-out resistance of max S = 1343 kN/m. For the
applied surface load of 500 kN/m, the limit state, therefore, could not be
expected. With the help of the measured strains (cf. figure 3.1), back-
calculation of the resultand tensile force leads to a value of approxima-
tely 100 kN/m. The measured (100 kN/m) and predicted (150 kN/m) tensile
forces are at least not inconsistent.

Instability because of external failure is predicted for a surface load of

1510 kN/m (cf. equ. (3.1), if the tensile strength is assumed for the ten-
sile forces in all reinforcement layers. This ultimate load for the retai-
ning wall, however, can never be reached, because for P~ 500 kN/m base
failure under this surface load occurs (cf. (15)).

For a suitable interpretation of the statics, therefore, the contradiction

between experiment and prediction can be removed. Consequently, rigid body
failure mechanisms are well suited for investigations of the stability of
geotextile reinforced retaining walls. The last statement. is also suppor-
ted by the experimental and theoretical results reported by (8).

An open question, however, is the separation of the resultan4tensile

force into the single reinforcement layers. From measurements of the
geotextile forces, corresponding to common service states, it is known
that the tensile forces are increasing from the bottom to the top of a
wall. Consequently, it can be expected that failure occurs first for the
upper reinforcement layers (lowest pull-out resistance). This, however,
should be incorporated in the formulation of safety requirements. This
problem has not been treated yet in theoretical or experimental investi-
gations.

3.2.2 Earth pressure distribution. The theoretical earth pressure

distribution incorporated in the presented design concept was chosen with
the help of the reported data from John (10) (cf. figure 2.6). More com-
plete published data from experiments, which could be used for verifica-
tion of the assumed distribution, are not yet available. The few
 329

additional data (cf. (1», however, indicate the following:

- Near the top of a wall the active pressure can be assumed,

- Near the bottom the earth pressures are much smaller than in the active
case.

The results of Berg (1), therefore, are consistent with the experimental
findings of John (10) (cf. chapter 2.1.2). For verification of the assumed
distribution more experimental data are needed.

3.2.3 Maximum tensile forces in the reinforcement. Very little experi-

mental data are available for the distribution of tensile forces in the
reinforcement (cf. table 3.1). In addition, the data reported by Wichter
et. al. (15) were gained from a large scale test with cohesive fill. The
presented design concept, however, is still restricted to cohesionless
fill materials. For the discussion of tensile forces, therefore, only'the
experimental findings from (16) can be used (cf. table 3.1).

These data were gained from a retaining wall of 7 m height with a steep
(1:0.2) face. The vertical distance of the golymer reinforcement layers is
~h = 1 m. Soil data: ¢ = 45°; ~= 14.4 kN/m • The measured stresses and
strains are due to gravity loads.

With the help of equ. (2.2) the following prediction for the maximum ten-
sile force can be made:
2
Z = 0.9. 14.4·7·1·tg (45°-45°/2) = 15.6 kN/m.

The maximum measured strain was 0.3 %. The corresponding tensile force can
be expected to be within the range 2.9 kN/m - 6.9 kN/m (cf. 16). A more
accurate back-calculation of forces is for several reasons (small strains,
inaccurate data for strain rates) not possible. The allowable tensile
force (tensile force divided by a factor of safety) is 31.4 kN/m; i.e. the
measured force is considerabley lower than the allowable tensile force.

With Z = 15.6 kN/m from equ. (2.2), the measured force is overestimated by
a factor of 2.3 - 5.3. Reduction of the friction angle - as is common
practice - by a partial factor of safety of 1.5 would lead to a overesti-
mation by a factor 4.2 - 10.1. At least for cases without surface loads it
can be expected that the predicted maximum forces are conservative.

For another large scale test (cf. (3» very low strain values «0.4 %)
are also reported, whereas these values correspond to common service
conditions. Under large surface loads (P = 500 kN/m) Wichter et. al. (15)
measured a maximum strain of 1.5 %. For the last mentioned test cohesive
fill material was used.

For all of these large scale tests the measured strains were markedly
smaller than the ultimate design strains, assumed to prevent creep. The
prediction of maximum tensile force, therefore, should be improved.

On the other hand, with the more sophisticated methods (for example FEM
with accurate constitutive equations for soils and geotextiles, including
nonlinear effects) needed for better prediction of stresses and strains,
only little practical experience has been made and/or the application is
expensive. Thus the application of more sophisticated methods cannot be
330

expected in the near future. For this reason the common design practice
should be improved with the help of experimental investigations. The aim
of such investigations should be a reduction of the vertical spacing of
the reinforcement layers. To attain this goal, reduction factors for the
calculated maximum tensil forces should be defined with the help of ex-
perimental data.

3.2.4 Distribution of vertical stresses. Experimental results for the

vertical pressures are reported by (3). In figure 3.4 the measured
vertical pressures are plotted versus the horizontal distance from the
wall face. 2 For the location of the pressure cells an overburden pressure
of 62 kN/m is reported, i.e. the pressure cells were located at depths of
3-4 m.
distance behind wall face

!
1
1
2I :3I 4I em)

FIGURE 3.4. Measured vertical pressures (Caroll and Richardson, (3)).

As mentioned before in section 2.1.2, in the vicinity of the wall face the
vertical pressures are much lower than the overburden pressure. Assuming
the length of the reinforcement to be 40 % of the height of the wall
(cf. (8)), for reasons of external stability, the reinforcement would end
1 - 1.2 m behind the wall face, and the vertical pressure over this area
is substantially lower than the overburden pressure (cf. ~ figure 3.4). This
effect is especially pronounced in the vicinity of the bottom of the wall.
Gudehus and Schwing (8), therefore, neglect the reinforcement layers, from
the bottom up to 10 % of the height of the wall, in stability calcula-
tions. This foregoing, therefore, is consistent with experimental data.

3.2.5 Displacements. As mentioned before, the conventional design prac-

tice makes use of empirical displacement estimates:

Horizontal displacements of the top of the wall in the order of magni-

tude of 1-2 % of the height of the wall after fade away of creep.

- By restriction of the design tensile force, as a certain fracture of the

tensile strength, fade away of creep can be assumed to occur within a
few months (cf. (16)). No prediction can yet be made for the elapse of
displacements versus time.
 331

For the discussed construction method the mentioned displacements of 1-2%

are not critical at all. Failure because of creep has not been reported
yet. On the other side, however, there is no experience for the behaviour
of reinforced retaining walls with respect to full service life (50-100
years) •

Gudehus and Schwing (8) simulated the case of an overstrained (8 %) rein-

forcement layer in model tests. For the displacements a linear increase
with respect to time was observed. Failure can be expected to occur very
soon. This model test result demonstrates the importance of adequate re-
strictions of creep deformations.

On the other hand, as already explained in detail in section 3.2.3, for

conventionally designed reinforcement layers the strains can be expected
to be less than 1.5 %; i.e. for such small strains creep is effectively
prevented (this holds at least for the short term experience with this
construction method).

To the best of our knowledge, long term displacement measurements (more

than 5 years) are not available yet. Meanwhile, Wichter (14) started
longterm displacement measurements for a retaining wall of 6 m height.
Initially displacements are registrated every week, then once every 2
months, and after a 2 years period once every 2 years. Parallel to the
displacement measurements, ageing of the reinforcement is investigated
with the help of so called control strips, embedded into the structure.
The described investigations are under work, results have not yet been
published .

3.3 Conclusions and proposals concerning future theoretical and

experimental work

The conclusions are presented in the sequence of the subsections of sec-

tion 3.2.

The experimental investigation of limit states should primarily be based

on model tests, as demonstrated by Gudehus and Schwing (8) for the stan-
dard retaining wall shown in figure 2.1. Additional model tests should be
performed for investigation of failure due to surface loads : For verifica-
tion (assumptions for the rules of similarity) of the model test results
only few large scale tests are needed.

Only little experimental data are available for the earth pressure distri-
bution. For verification of assumed distributions, therefore, additional
data from large scale tests are needed. Force measurements in connectors
are best suited for investigations of the earth pressure distribution.

For the maximum tensile force in a reinforcement layer, corresponding to

the limit state of external stability, no experimental data are available
yet. In p~rticular with the help of testing results the problem could be
solved, how the resulting tensile force (derived from a limit state equa-
tion) is distributed into the single reinforcement layers. Also conse-
quencesfor the statement of safety requirements from such measurements
can be expected (reinforcement pulled out collectively?; consequences re-
sulting from failure of a single reinforcement layer?).
332

For the maximum tensile force, corresponding to internal stability, test

results are available. The measured forces are overestimated by the pre-
diction. Consequently, the number of necessary reinforcement layers is
overestimated, leading to higher expenses. With the help of available
experimental data and additional large scale tests correction factors for
the predicted maximum tensile force could be defined.

For predicting the pullout resistance, good knowledge of the vertical

stresses is essential. The few experimental data to this subject indicate
that the vertical stresses, in particular in the vicinity of the bottom of
the wall, are essentially lower than the overburden stress. For a better
understanding of this problem, more experimental data from large scale
tests are needed.

Displacement measurements in particular should be performed in the long

term range with respect to the questions concerning ageing of geotextiles.

For the derivation of partial safety factors, more theoretical and experi-
mental work has to be done:

- Basic statistic data from experiments (pull-out resistance),

- Back~calculation of the reliability of well performing retainig struc-

tures with the aim of calibration of the applied statistical methods,

- Partial safety factors for loads and material strengths.

4. CASE STUDIES

For the following case studies, 3 typical retaining walls, all located in
Switzerland, have been selected.

For construction of a new communication road between Chiasso and

St. Gotthard, a temporarily road had to be built on a steep hillside. As
solution, a geotextile reinforced retaining wall was chosen (cf. figure
4.1). Design was based on internal stability (assumption of active earth
pressure). Under the calculated tensile forces the design~trains of the
reinforcement were less than 5 %. Vertical displacements have been mea-
sured over a 2 years range (cf. (11)). The largest displacement was 18 mm.

For construction of a parking lot in St. Gallen, the retaining wall from
figure 4.2 was chosen as construction method (12). The greatest height of
the wall is 6 m.

The retaining wall from figure 4.3 was used for construction of a new
railroad track (9). The geotextile reinforced wall was designed similar as
an anchored wall. For more details concerning the fill material and the
reinforcement see figure 4.3. The wall is not instrumented.
 333

Colbond P450
Sodoco AS 420

cO.7m

a.)

b)
FIGURE 4.1. Temporary road: a) Cross section, b) Road after construction

WendeplatZ

- - - " - - " - ' - - <;;

height 6 m .'· ....... t ..

FIGURE 4.2. Parking lot

 w
Auflast : 1 Gelelse w
= 2 x B to/ml Ruf ca. ~ m Breit e -I>-
; }.2 to/m'.

Itt. I 1'-'.'
;C··..
.t:- ' ..0.. ',~. '~',
• •' ~"" ~
• •''''D ".•.", o · ~ .~./
~ N r. 17 v 387.44 -t -I I _ . . ·
evtl. Dlchtungsbahn und DraJnage I eltung
2-3 ~ / zu Oelabscheider :
---.. .. i' - chemisch. resistent
verschweisste Fugen
hohe Dehnbarkeit und Festigke it
hohe Kerbfestigkeit
Wurzel fest
Nagetier-resistent
n I
arbor, I\'

~
Typ HaTe Gewebe Nr. D 00.5}0
Reisakraft longs 51.0 kN/m
Oberlappungen min. 0.30 m
Bahnen zu 7.65 m Liinge
oberate 2 Bohnen 5 + S + 0.60 :. 10.6m
J (Nr. 16 und 17)
Gesamtausmass ca. 13'000 m 2

SchGttmateriai
wtrdichttfer
I siltiger sandiger Kies
in Legen zu max. 40 em im
Endzustand
VerdichllKlg mit schwerer 5 to Vi bra-
tionswalze auf max. 95% Proctor
5 min 90% Proctor in der Randpartie
im Bereich 1 m und rnehr vorn Rand
ME min. = 600 kg/em'

,
,...-ce. Aushublinie

Lagenweise aufgefUJlt
2.UG und verdichtet

FIGURE 4.3. Cross secti on of geotextile reinforc ed r e tai ning wall

 335

5. SUMMARY

In this paper mainly the knowledge, which can be gained from large scale
tests for the design of geotextile reinforced retaining walls, is dis-
cussed. This discussion is based on published design concepts and instru-
mented data. Proposals for future theoretical and experimental work are
made.

The discussion of instrumented data is based on a design-proposal pub-

lished by Gudehus and Schwing (8). The limit analysis design incorporated
in this concept is verified by model tests. Partial factors of safety have
been derived by the above mentioned authors by application of probability
methods. This concept holds for cohesion less fill materials. Concentrated
surface loads (strip loads) are not yet incorporated within this concept.

On the basis of the above briefly explained design concept, the evaluation
and interpretation of instrumented data is discussed and applied on pub-
lished data. The following results were derived:

- Ultimate loads, derived with the help of rigid body failure mechanisms,
are consistent with observations from large scale tests. Reported incon-
sistents between predicted and measured loads (cf. (15)) can be shown to
be a result of a static misinterpretation.

- For investigations of limit states model tests are better suited than
large scale tests.

- The measured tensile forces (of the reinforcement) are overestimated by

the prediction.

- As suggested by Gudehus and Schwing (8), the vertical stresses, in

particular in the vicinity of the bottom of a wall, are remarkably
smaller than the overburden pressure. This should be considered for
estimates concerning the pull-out resistance.

The available instrumented data from large scale trials are insufficient
for verification of the design concept. The following additional data are
needed:

Distribution of maximum tensile forces (with respect to depth) for the

reinforcement layers in limit states (external stability),

- Earth pressure distribution,

- Maximum tensile forces corresponding to internal stability,

- Distribution of vertical stresses (with respect to depth),

- Long term displacement measurements (ageing of geotextiles),

- Basic statistic data, for example distribution and corresponding

parameters for the pull-out resistance.
336

Future theoretical work should be concentrated on the formulation of

safety requirements (partial factors of safety) needed for application of
limit analysis design.

REFERENCES

1. Berg, R.R., Bonaparte, R., Anderson, R.P. and Chouery, V.E.: Design
construction and performance of two geogrid reinforced soil retaining
walls. Proceedings of the Third International Conference on Geotex-
tiles, Vienna, Austria, 1986.

2. Breitschaft, G. and Hanisch, J.: Neues Sicherheitskonzept im Bauwesen

aufgrund wahrscheinlichkeitstheoretischer Ueberlegungen - Folgerungen
fUr den Grundbau unter Einbeziehung der Probennahme und der Versuchs-
auswertung. Proceedings of the Baugrundtagung, Berlin, 1980.

3. Caroll R.G. and Richardson, G.N .• : Geosynthetic reinforced retaining

walls. Proceedings of the Third International Conference on Geotex-
tiles, Vienna, Austria, 1986.

4. Dutch/Enka Method. Calculation method to determine the internal stabi-

lity of reinforced earth structures. Unpublished report of the Enka
Industrial System, lEN 2.297 Ris/At., 1985.

5. Geotextilhandbuch. Published by Chemie Linz AG. Austria, 1986.

6. Go1dscheider M.: Standsicherhei tsnachweis mit zusamme"ngesetzten Starr-

korper-Bruchmechanismen. Geotechnik 2, 130-139, 1979

7. Gudehus, G. and GassIer, G.: Soil-Nailing - Some aspects of a new

technique, ICSMFE X, Proc. Vol. 3, Sess 12, Stockholm, 665-670, 1981

8. Gudehus, G. and Schwing, E.: Standsicherheit kunststoffbewehrter Erd-

bauwerke an GelandesprUngen. Proceedings of the Baugrundtagung,
NUrnberg, 1986.

9. Jaecklin, P.: Unpublished data - Jaecklin Consulting Eng., Ennetbaden,

Switzerland, 1986.

10. John, N.W.M.: Geotextile reinforced soil walls in a tidal environment,

Third International Conference on Geotextiles, Vienna, Austria, 1986.

11. v. Krannichfeldt: Unpublished data. Ufficio Strade Nazionali, Ticino,

Switzerland, 1986.

12. RUegger, R.: Unpublished data. RUegger AG, Consulting Eng.,

St. Gallen, Switzerland, 1986.

13. Werner, G. and Resl, S.: Stabilitatsmechanismen in geotextilverstark-

ten ErdstUtzkonstruktionen. Third International Conference on Geo-
textiles, Vienna, Austria, 1986.

14. Wichter, L.: Geotextil-Erde-StUtzwand als Dauerbauwerk. Bautechnik 62,

289-291, 1985.
 337

15. Wichter, L. Risseeuw, P. and Gay, G.: Grossversuch zum Tragverha1ten

einer Steilwand aus Gewebe und Mergel. Proceedings of the Third Inter-
national Conference on Geotextiles, Vienna, Austria, 1986.

16. Yamanouchi, T., Fukuda, N. and Ikegami, M.: Design and techniques of
steep reinforced embankements without edge supportings. Third Inter-
national Conference on Geotexti1es, Vienna, Austria, 1986.

17. van Zanten, V.R.: Geotexti1es and Geomembranes in Civil Engineering.

 REVIEW OF SESSION

CASE HISTORIES AND FULL SCALE TRIALS

Data obtained from instrumented full scale polymer reinforced structures

are of immense value in understanding the behaviour and verifying the
design concepts for such structures. However, there are two major diffi-
culties facing researchers, firstly shortage of funds for full instrumen-
tation of trial structures and secondly lack of an agreed list of relevant
case history data and a common format for their presentation. Consequent-
ly, as Christopher pointed out, of an estimated 200 slopes and walls con-
structed in the USA and Canada using polymeric materials, data on only 46
projects could be collected. Of these only 6 projects were fully instru-
mented. He presented these data in four comprehensive tables covering
general information; design information; actual properties of the rein-
forcements and design methodology; and project instrumentation and moni-
toring information. This work represents a substantial step forward and
workers in this field should be encouraged to extend such information so
that a comprehensive data bank of case histories is constructed. It is
important however, that when consulting such data the user is aware of all
the parameters and assumptions used in the design.

Delmas has provided valuable data on full scale geotextile reinforced

structures in an attempt to determine the influence on their behaviour of
a) the method of placement of the formwork and the facing; b) the use of
clayey soil and c) the long term behaviour of the geotextile. Five
instrumented structures were described in detail. The results illus-
trated, perhaps not surprisingly, that the tensions in the reinforcing
sheets were significantly influenced by the placement method and by the
nature and timing of the stresses to which the structure is ~ubjected. He
recommended the use of placement methods which not only facilitate the
building of the structure but also ensure the optimum tensioning of the
geotextile. By monitoring the pore water pressures in two structures he
concluded that the use of soils having a large percentage of fines posed
no special problems provided that the placement and compaction do not
cause build up of pressure and the geotextile is capable of dissipating
any increase in the pore water pressure. This is important to ensure
optimum soil-geotextile friction. It was interesting to learn that ageing
of geotextiles, which was examined by testing samples 15 years after
placement, caused loss in strength and in this particular case a gain in
stiffness.

Studer emphasised the importance of instrumented large scale tests in

selecting appropriate design procedures and safety factors, he highlighted
the lack of reliable comprehensive data in the literature and presented a
detailed list of the required data. He argued however, that large scale
tests are not suitable for the determination of the critical geometry and

339

P . M .Jarrett and A. McGown (eds.), The Application ofPolymeric Reinforcement in Soil Retaining Structures, 339-340.
© 1988 by Kluwer Academic Publishers.
340

load and that model testing should be used for the systematic
investigation of the behaviour of reinforced walls in limit state and the
determination of the failure mechanism. It is interesting to note that
although the measured vertical stresses at the base in the vicinity of the
wall were significantly smaller than the overburden pressure, it is
accepted that the average value of the vertical stress at the base is
equal to the overburden pressure.

Leflaive presented an interesting contribution regarding full size testing

of Texsol walls. It might be helpful to point out that Texsol is a
mixture of synthetic fibres and granular soil prepared in the field using
special equipment. The walls were 3 m high and surcharged up to failure.
The main and intriguing conclusions from these tests were that the failure
was not characterised by a shear plane but occurred by the walls
overturning and that the deformations were extremely small up to almost
90% of the failure load. It will be interesting to see the results of the
planned centrifuge tests on this material. Jones pointed out certain
geometric similarity between this technique and that used by the
Victorians of placing stones on steep slopes. After 100 years or so,
which is a considerable period of time, these stones needed replacement.

Leflaive drew attention to a very interesting observation regarding the

determination of the factor of safety. It was found that when a yarn of
the polymer was statically loaded with 60% of its failure load it fails in
about two days. On the other hand, a Texsol sample loaded with 60% of its
failure load for one year did not fail. This implies that the yarns in
the Texsol sample are not loaded at 60% of their failure load but reach
this value only when the load on the sample is close to the failure load.
Although these findings seem consistent with the fact that strains in
Texsol are very low prior to failure, they make the determination of the
factor of safety quite difficult.

Berg gave a detailed review of the design, construction and performance of

a5.3 m high and 99 m long polymer-geogrid reinforced soil seawall. The
wall is exposed to severe climatic conditions involving wave action and
freeze-thaw cycles. Two sections of the wall were instrumented in order
to monitor movements of the wall face and the reinforced soil mass. In
addition attempts were made to measure the tensile loads in the geogrid
layers. This is an extremely valuable example of a well documented case
history and Mr. Berg's plan to continue monitoring the performance of the
wall in future will certainly add to its value.

Scott reported on a large fully instrumented trial embankment to be

constructed in Devon, Alberta in the summer of 1987. A series of sections
are to be reinforced with various grids in an embankment 12 m high built
from cohesive fill at a slope of 1:1.

In conclusion, this session has demonstrated the value of full scale

trials and highlighted the difficulties encountered in monitoring their
performance. It also emphasized the urgent need for developing standard
guidelines on testing methodology and data collection. This information
should be well documented in a comprehensive data bank of case histories.
Evaluation of Material Properties
THE PURPOSE OF MATERIALS EVALUATION AND RED:MMENDATIONS FOR
MATERIALS TESTING

E. LEFLAIVE, Laboratoire Central des Ponts et Chaussees, France

A. MCn::WN, University of Strathclyde, scotland

INTRODUCTION
Reinforced soil wall structures comprise a number of components and
materials; the soil backfill, reinforcements, facing units, joints and
connections, foundation for the facing units and the subsoil on which the
structure is built. All are important to the efficient performance of the
structure and should be properly evaluated by appropriate testing.
However, it is the use of polymeric reinforcements which is the specific
point of interest for this Workshop thus this paper deals only with the
evaluation and testing of these. Further, detailed descripticns of test
methods are not included as they are not considered to be the points at
issue, rather it is the choice of which test methods that should be used
that is critical.
The report is written in two parts: the first deals with the purpose of
materials evaluation -and the second provides a recorrmended approach to
materials testing.

PART A: BACKGROUND PURPOSE OF MA'IERIALS EVALUATION

For soil retaining structures reinforced by geotextiles the question of

evaluation of the reinforcing materials has to be examined with reference
to three different aspects: design concepts, constructioQ. practices and
testing requirements for other end uses.

1. Design Concepts
The design of any reinforced soil retaining structure requires the study
of both its external and internal stability.
For the purposes of external stability calculations, the retaining
stucture is generally considered to be a rigid body, therefore only tJ:e
properties of the surrounding soil are genera 11 y required to be known,
apart that is fran the mass (weight) of the structure. There is, hcwever,
a growing awareness that the overall deformability of the structure will
influence such factors as base pressure distributions and lateral earth
pressures and that these may in turn influence ttE external stability of
the structure. In such cases the properties of the soil and reinforcement
forming the stucture may have to be taken into account. This is not
343

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 343-355.
© 1988 by Kluwer Academic Publishers.
344

current practice and is only mentioned here to point out the possiblity of
future developments. Thus external stability will not be discussed further
in this report.
In contrast, a full knowledge of all the materials properties is
certainly necessary for the calculation of the internal stability of
reinforced soil retaining structures. The specific data required for the
reinforcing of materials depends very much on the analytical approaches
taken. These may be classified as follows:

1.1 Discrete Constituent Materials Approach

The principle of this approach is that an analytical model of the system
is established within which discrete sets of test data are used for the in-
isolation properties of the reinforcements and for their interface
interaction characteristics with the soil.
This philosophy is quite cornmon in engineering. In the field of
geotextiles, it has been very clearly expressed by Studer (1982) on behalf
of a Swiss group, at the Second International Conference on Geotextiles:
"The basic idea of this concept is as follavs: if the properties
of the soil around the geotextile are well known and if a model of the
soil-geotextile interaction does exist it is always possible to investigate
the behaviour of the soil-geotextile system analytically at a later stage.
Thus an analogous concept like that used for composite structural
materials, (e.g. reinforced concrete), is applied. However, in the field
of geotextiles, analytical models are still at an elementary stage.
Nonetheless, it is the committee's belief that this is the only physically
correct procedure from an engineering point of view".
It follows from this approach that all specifications for geotextiles
used as reinforcements should:
"rely on standard tests on the geotextile only, in spite of the
fact that a geotextile is never used as a pure construction material but
always in conjunction with soil."
It may be noted that, strictly speaking, frictional characteristics have
to be measured with the reinforcement in contact with another material,
whether it be a soil or any other frictional surface; but as long as the
friction test remains a standard test with standard components the general
approach is not modified.
This paper from Switzerland was presented at a Session 2B of the Second
Internation Conference on Geotextiles, Las Vegas, (1982), which was devoted
to International Standards. Apart from one paper from Finland, dealing
only with the use of non-woven geotextiles in road construction, and one
Dutch paper on vertical drain specifications, six other papers gave general
considerations to the specification and testing of geotextiles as discrete
constituent materials and listed recc:mrended testing procedures. These
papers were from U.K., Germany (BRD), France, U.S.A., and from two
international societies; PIARC and RILEM.
 345

The German, French, PIARC and RILEM papers essentially sUpJX>rted the
Swiss paper and presented testing methods only for geotextiles in-
isolation. However, both the U.K. and U.S.A. papers mentioned that in sane
circumstances a choice rray have to be made between data fran in-soil and
in-isolation testing. The paper from U.K. highlighted this to the greater
extent.

1.2 Equivalent Composite Materials Approach

The principle of this approach is to replace the system formed by the
soil and its reinforcement by an equivalent homogeneous material. In
practice it can be applied only in situations where the distance between
reinforcing elements is very srrall corrpared to the retaining structure
dimensions, in other words, the scale at which the canposite is formed has
a different order of magnitude from that of the structure. In geotextile
technology, "Texsol" and "Mesh Elements" systems have been proposed that
satify this condition and are or could be used for retaining structures.
Such systems have not been included in the testing program associated with
the Workshop, nevertheless they are an application of polymeric
reinforcement in soil retaining structures.
Soil-reinforcement interactions (friction, local stress effects, structural
interactions, etc.) are autanatically incorporated in the testing employed
for the equivalent canposite materials approach. For this reason, this
approach, when it can be applied, is the simplest one. Indeed, it is
nothing but a common application of mechanics of solids to real materials,
albeit that the composite rraterials are neither continuous nor homogeneous.

1.3 System Simulation Approach

A third approach to the design is the system simulation approach. In
some ways it can be considered as an intermediate solution to be used when
both the discrete constituent and the equivalent composite materials
methods are neither applicable nor available. The objective of this
approach is to simulate the system behaviour, in part or as a whole, and
to use it to predict the behaviour of a family of structures having in
corrmon a number of features with the simulated system. Again, it is a very
common design approach when engineers have to give answers to complex
problems where all the phenomena involved are not fully understood and
cannot be easily measured se?arately.
In practice simulation testing for reinforced soil structures may be
carried out in two ways. One is to test the reinforcments in the
laboratory when incorporated into the soil and stress environments in which
they will be used in the structure. This is done in order to obtain
results that are more directly representative of their actual operational
behaviour. The other is to test prototype structures, either at full size
or at model scale. These model tests must not be confused with research
work on models as it is intended that these will provide data which can be
346

used directly in the design of an actual project.

'!he system simulation approach was strongly re=rrmended by Murray and
McGown (1982). In this paper, geotextile tests are divided into two
groups: Index Tests and Design Data Tests. In-isolation tests and in-soil
tests using standard soils, form the index tests group, which are
considered to be valid only for quality control, (during and after
manufacture), and to have little validity for design purposes. Design data
tests it is suggested can be obtained from in-soil tests using the soil of
the project and from prototype trials. However, it is recognised that
developments and experience may make possible the use of index test results
in design so that in the future, the differences between this simulation
approach and the discrete constituent materials approach may greatly
reduce.

2. Construction practices
The point here is that the analytical approaches used in the design
process soould be vie\\ed only as the theoretical operation leading to an
overall understanding of the stabi l i ty of tiE structure and to its basic
dimensioning. In contrast, practical engineering requires that what
happens during construction must be taken into account as it may affect the
resulting engineered structure in many ways, among which the material
properties are an important aspect. Therefore materials evaluation must
indicate to what extent reinforcing elements may be modified or damaged by
planned or unexpected events during the building of the structure.
'!hus, material testing must give the engineer information on possible
material modifications due to construction procedures, and the tests
required to do so may be quite different from those envisaged for
calculation needs. Theoreticians may sometimes consider these tests to be
of minor importance and include into the Factor of Safety such
uncertainties on the actual material performance. HO\\ever, experience
shows that such material modifications must be fully assessed and
independently allowed for. Failures in structures have often resulted from
material damage due to poor field workmanship or unexpected building
conditions.
Once in the ground the reinforcements may tiEn suffer mechanical and
chemical degradation. For geotextile reinforcements the most obvious risks
are mechanical degradation due to accidental tensioning or tearing,
puncturing and abrasion. In same cases a possible reduction of performance
may also result from loss of permeability due to clogging where the
reinfor~ement is also expected to have a draining effect. Problems related
to joints and connections may also have to be considered from the
construction point of view. It is also possible that degradation may
result from changes imposed on the polymeric constituents of the
reinforcement due to temperature changes, chemical and biological attacks,
exposure to light, fire or even animal attacks. When buried in most soil
 347

structures the resistance to degradation from all of these agencies is

considered to be very high but the need to develop test rrethcds to justify
this ccnfidence must be one of the ITDst urgent requirEments in this field
of study.
Another question to be discussed, relates to the behaviour of the
reinforcements if they are not laid within geometrical tolerances.
Avoiding waves or folds leading to a "Slack" effect may be viewed
essentially as an inspection problem, but it may also be related to
material properties and so special tests (e.g. flexibility). The
possibility that a geotextile can conform to a very uneven surface can be
measured by flexibility tests to a certain extent but more fundamentally by
evaluating its ability to change its area during different types of tensile
tests. This leads to an interesting question, viz., should the designer
look for a fairly rigid reinforcing material to obtain a good laying
geometry or for a geotextile that can easily conform to an uneven surface,
to insure good contact and so interfacial friction conditions? This choice
may need to be clarified by simulation testing. It is unlikely that
standardised index testing will assist in this.

3. Testing Requirements of other End Uses

Testing rrethcds must be appropriate for the particular application of
geotextiles under consideration, i.e. soil reinforcerrent. But geotextiles
are also used for other purposes and it would be unreasonable and
inefficient in practice to devise independent! y as many sets of testing
rrethcds as there are particular applications. Since properties that are
useful for reinforcement design are also of interest in many other
applications, although with different degrees of importance, it may be that
a measure of compromise will have to be reached between the various field
of application to limit the number of testing rrethcds adopted. However, it
is realised that this is not an easy gcal to achieve beca~se many aspects
have to be taken into account. Most certainly, potential users of these
materials must not be discouraged by an overcomplexi ty of test methcds.
Nonetheless, testing should be rigorous (i.e. time consuming or expensive),
if its aim is to determine once and for all the technical characteristics
of a given product. Whereas, it must not be time consuming or expensive if
it is for identification purposes only. If tests have to be performed for
each project then this is where a compromise must be found.
To avoid confusion between the different types of tests, definitions
proposed by RILEM (1985) are restated here:

"Test can be classified in 3 categories:

- control tests in factories
- reception or identification tests
- suitability tests."
348

"A control test in a factory is a means of

following the production of a geotextile. The
manufacturer is responsible for the choice of his
rrethod.
The results of these controls are confidential.
External organisations can interact in two
different ways with these procedures:
1) to recommend the control of certain
parameters because they appear to be significant.
(This would simply be in the form of a
recorrmendation) .
2) to establish agreerrent procedures or quality
assessment certification to verify that a
manufacturer possesses all the necessary
manufacturing knowhow and the material testing
facilites for quality control."
"A reception or identification test is devoted to
the identification of a product by the end user
and involves identification of:
- raw material;
- manufacturing procedure;
- physical properties (components and final product).
These characteristics are given by the
Manufacturers and are their responsibility. They
may be best presented on technical cards and
include data which can be very simply verified by
Contractors and other end users. Obviousl y there
must be a good agreement between Manufacturers and
Contracters regarding the tests procedures
employed to obtain these data."
"A suitability test is devoted to the selection of
a product for a given end use. The test to be
chosen depends on:
- the function(s) to be employed
- the analytical procedure to be used.
It must be noted here that this classification is
independent of the different analytical approaches
mentioned previously. It does not tell whether a
suitability test is a simulation test, an in-
isolation test or a test on the composite. Its
usefulness is to avoid misunderstandings due to
possible confusion between factory control,
commercial and technical identification and
description, and suitability evaluation for the
design of a project."
 349

An additional factor is that data may require to be presented in

different fashions for different applications, for example:
(i) Information may be presented quantitatively (e.g. mass per unit area =
250 g/m2) or qualitatively (e.g. the geotextile is made of polyester).
The first type involves a measurement; the second requires a test to
identify a quality.
(ii) Information may also be given by a test which may be quite specific or
has a wide range of applications; in that sense a test may have an
extensive value or be specific, a point discussed previously.
(iii) In view of a given function or behaviour, information given by a test
may be directly applicable for a classification or a calculation, may
have an indirect significance (e.g. whether the geotextile is woven or
non-woven may be useful for interpreting tensile and friction test
results), or may have no direct significance; in that case it still
may remain useful to validate other data by cross-referencing.
This last point is irrportant for the reason that good design and good
material evaluation must reduce to a minimum the probability of making a
mistake in the choice of materials. In this sense, some data may be useful
in spite of the fact that they do not directly fit into a calculation
rrethod; therefore these data must not be neglected and must be used as long
as they are easily obtained.

4. Swrmary
It may be stated that data on materials may interact in many different
ways within the design process and that material evaluation must give to
the Engineer sufficient information to enable him to:
a) check that all the data available are consistent and, thus, reduce the
probability of an error on material identification and properties.
b) refer to existing projects built with similar materials and, thus,
make use of experience gained in the past.
c) dispose of all data required by the reinforcement dimensioning
method(s) to be used.
d) assess possible material modification, degradation and/or misuse due
to construction conditions, and evaluate potential construction
problems.

PART B: RECOMMENDED APPROACH TO MATERIALS TESTING

1. Identification

Identifying a number of possibly suitable reinforcing materials is the

first step in the process of selecting a product for a given project. The
identification of these products should include a canrrerical reference,
name and number, a description of the product and a list of basic physical
characteristics. The purpose of this, as mentioned previously, is to
350

enable the Engineer to check that the data available is consistent with
previously published data and to allow reference to past experience.
Identification testing is best when it is rapid and easy to control, while
giving the experienced Engineer a good picture of the product and allowing
an assessment of its overall performance with sufficient accuracy for first
stage design.

Identification testing for geotextile 'reinforcements should provide data on

properties such as those sh= in Table 1.

TABLE 1 IDENTIFICATION DATA FDR GEDTEXTILE REINFDRCEMENTS

Mass per unit area 1

Nominal Thickness (or other relevant dimensions) 1

Tensile Strength and Strain to Failure (for method 1

see Section 2) 1
Puncture Resistance 1

Tear Resistance 1

Abrasion Resistance 1
______________________________________ 1

2. Reinforcement dimensioning data

There are two main groups of characteristics:

- tensile properties (short and long terin loading)
friction (or soil-geotextile interaction) properties.

2.1 Tensile Properties

2.1.1 Short Term Loading
There is a general agreement that the unidirectional constant rate of
strain tensile test is the most appropriate test methodology to determine
the short term tensi Ie behaviour of geotextiles. A number of specific
testing methods have been devised to approach as closely as possible the
performance of a geotextile strained unidirectioIlally, i.e. withe 2 = o.
The different methods applicable to geotextiles above vary according to:
- the system for the application of stress and strain
- the size and shape of sample
- the way of expressing test results.
Three main systems have been proposed for the application of stress and
strain: - cylindrical deformation by air or fluid pressure
- biaxial testing on cross-shaped sample
uni-axial testing.
The first two systems are usually considered to be too cumbersa:ne for
corrmerical testing purposes. They are, however, of great interest for
 351

research testing because they make possible testing conditions difficult or

impossible to obtain otherwise. Also the stresses and strains can be
established in both principal directions, which is of basic importance for
theoretical interpretation. On the other hand, they have sane technical
drawbacks, which are independent of their mechanical complexity, as
folla.vs:
Firstl y , when cylindrical deformations are obtained on a rectangular
sample submitted to a pressure on one side or by inflating a sleeve then:
there is a difference of pressure between the two faces of the
geotextile which is related to the strength of the material, the
curvature of the sample during the test, sample size and geotextile
deformability. This gradient of 0""3 cannot be avoided. On the other
hand this type of system can easily be adapted to apply an additional
pressure in the direction of the thickness of the material.
if the deformation of the sample is close to cylindrical, E 2 is small,
but remains usually positive, contrary to the uniaxial testing where
E2 is negative.
Secondly, although plane biaxial testing allows control of stresses and
strains in the planar condition, it does not normally reach failure. These
tests are therefore more of interest for modulus determination than for
ultimate strength evaluation.
Thirdly, although testing with stress and strain measurement in both
directions is necessary to study the mechanical behaviour of planar
structures, little is known about the biaxial behaviour of such materials,
particularly the effect of manufacturing technique on the interrelation
between stresses and strains in the different directions of the plane. For
example woven products are often considered to be highly anisotropic but it
seems that this is not so much so when biaxial stresses are applied.
Theoretical models of such orthotropic structures with tensile resistance
only are not available. Even if one considers only ~resistance to
deformation in the plane, neglecting bending and torsion, theory would be
useful to correlate different testing conditions, such as tensile tests
with different values of £2' and to analyse the relationships between
traction, shear and area change.
As mentioned previously, present practice for geotextile reinforcerrent
testing is to consider only uni -axial deformation and to approach this
situation experimentally. This is done most simply by traction on
rectangular samples in conditions such that E 2 is small. In order to
reduceE 2 in non-woven products to a minimum, one solution is to use
transverse" brackets with pins opposing restraint of the material during
testing. This system was developed by the University of Strathclyde and
first reported on by Sissons (1977) for 200 x 200 rom samples and is used in
Switzerland with 100 (width) x 200 (length) rom samples. Another approach
is to confine the geotextile in-soil, but the effect of soil confinerrent
has influences other than opposing lateral restraint. This mode of testing
352

is discussed in a later section.

'!he most conm:m solution for reducing the influence of E. 2 is to adopt a
wide width sarrple, i.e. with width larger than the length. This approach
is favoured by many engineers and researchers. Discussions on this test
concentrate on the following points:
- width/ length ratlo
rate of strain
use of measured C2 values for correctingfl values

It is very often stated that a width/length ratio of 2 (200 x 100 mm

samples) is sufficient for most geotextiles, including non-WQvens. '!he
rrain advantage of using this is that i t limits the force that the equipnent
must develop for strong materials, (as a first approximation ·forc e is
proportional to sarrple width). However, it is considered by others that it
is necessary to adopt a larger of width to length ratio say of 5,
associated with a correction taking into account the measured value of E 2
to approach the ideal situation of £2 = o.
Applying this correction is equivalent to determining the area change of
the sample during the test. It is found that the maximum area change is
larger with a width/length ratio of 5 than with a ratio of 2. Corrected
values of tl are srraller than measured £1 values. Their significance is
that they correspond to the maximum area change possible with the material
and that shear deforrration resulting in lateral restraint is compensated
for.
In tests with width/ length ratio of 2, measured values of £1 at a given
percentage of failure, tend to be larger than with a width/length ratio of
5 because a larger shear deforrration occurs but, at the same time, tends to
be srraller because area change is smaller. '!hus, these values are not
really meaningful for geotextiles where large C 2 values develop during
traction. The£2 correction is not normally applie,? to test results
obtained on samples with a width/length ratio of 2. In general,
width/length strength values found with a ratio of 5, are 5 to 10% larger
than with a ratio of 2 for non woven geotextiles.
It has been shown that for needle punched non-wovens that by using test
specimens with a width/length ratio of 5 and making corrections to the
measured values of E. l' the load-strain behaviour obtained is very similar
to that from in-soil testing with a width/length ratio of 2, a confining
stress of 100 kN/m2 and sand as the confining medium. However, the effect
of soil confinement is known to vary with the soil type and applied stress
level, · thus no unique relationship exists between these two test methods.
'!he in-soil test method, like the tests using width sarrples with a length
ratio of 5, is most applicable to needle punched non-woven geotextiles.
In-soil testing on melt-bonded non-wovens produces data which is only a
little different from that in-isolation, (the higher the degree of bonding,
the smaller the effect), and there is almost no difference for woven
 353

geotextiles and grids or nets whose structure is aligned along and across
the line of traction,although sample shape has same effect on elongation at
failure.
The constant rate of strain at which the short term tests are carried out
does vary from Country to Country but it is becaning more standardised and
is most often now carried out at 10 per cent per minute. However, a recent
proposal is that the test should bring the test specirren to failure in 2
minutes ~ 10%. These rates of strain are too high for manual recording of
test data and so automatic data recording is generally required.
The effect of the rate of strain during testing on the measured load-
strain behaviour of geotextiles varies with both their polyrrer content and
their structure. The behaviour of wovens, girds and nets is predominantly
controlled by their polyrrer content and the effect of the rate of strain is
significant whereas needle punched (and to a lesser extent melt-bonded)
products, being structure controlled are relatively less sensitive to rate
of strain, particularly at lOt.' strain levels. For the same reasons the
effect of the temperature of testing is significant for wovens, grids and
nets and less important for non-wovens.
2.1.2. Long Term Loading
Once again there is general agreement that unidirectional tensile tests
should be used to determine the long term tensile behaviour of geotextiles,
however, only a limited amount of long term testing has been carried out.
What testing has been done has involved sustained constant loading at
constant temperature on a number of different test specimens. The same
discussion on sample width/length ratio and in-isolation to in-soil
testing, applies to long-term testing as much as to short term testing and
as before the most commonly adopted width/length ratio is 2, although for
woven geotextiles much narrower test specimens are quite often used. Very
few long term in-soil tests have been undertaken. Nevertheless, such tests
on soil-textile composites as have been performed, show that a very
different creep behaviour of the canposite is obtained canpared to the
reinforcing element in-isolation.
Just as important as the test methodology for these long term tests is
the rrethod of presenting the test data. Very often the data is plotted as
strain versus the log of time and a creep coefficient derived fran the
slope of the linear part of the plotted data. This requires that the
initial non-linear part of the plot is ignored, however, this may represent
the majority of the strains developed in the geotextile at that load level.
Thus a material which exhibits very large strains initially but little
strain thereafter, may appear to be a more eff icient soi 1 reinforcement
than a material which has only limited initial strains and modest time
dependent strains. For this reason the total strain from the time of
loading should be measured and included in data presentation. A test
rrethodology and rrethod of data presentation which accanplishes this has
been developed in U.K. between the University of Strathclyde and TRRL,
354

Murray and McGown (1987). It is this method which has been used to present
the data for use in the predictions of the test walls which form the
background to this Workshop.
Only by carrying out such long term tests can the long term tensile
behaviour be correctly established but this type of testing is very time
consuming and expensive. Thus as discussed in Part A it is not expected
that this will be undertaken for routine purposes. Consequently it is most
important that the data obtained from such testing be correlated very
closely with the short term test data which will be routinely obtained in
order that checks on the quality of materials can be properly established.
Recent work suggests that checks on the short term ultimate strength are
not a sufficient measure of material behaviour and that it is the entire
load-strain curve that must be considered. This should be the subject of
future stud ies.

2.2. Soil-Reinforcement Interaction

This is usually achieved by the development at the interface between the
soil and the geotextile, but in certain products such as grids and nets
interlocking between the soil particles may also develop. Two conflicting
approaches exist for the determination of these interaction properties i
pullout tests and direct shear tests, the choice of which test method to
be adopted is often dictated in design methods. Much more work requires to
be undertaken to relate these two test methods and to standardise on the
specific test apparatus and techniques. Further the influence of such
factors as soil type, canpaction levels and methods, and reinforcements
stiffness needs to be more fully investigated. lastly, investigation of
the influence of redistribution of stresses at the soil/reinforcement
interface with time as the soil strains or as the reinforcement creeps or
degrades, needs to be assessed.

3. Other Properties
Other properties which may be of critical significance to the long term
operational behaviour of reinforcements may be discussed in two groups as
follows:

3.1. Durability
The long term degradation of geotextiles is, as stated previously, one of
the most obvious risks to their satisfactory long term behaviour.
Unfortunately there is almost no agreement on suitable test methods to
establi's h this behaviour. This does not mean that geotextiles will not
perform satisfactorily or that many of the available, but unsuitable, tests
should be adopted until suitable tests are developed. Rather it means that
there is an urgent requirEment for this aspect of materials evaluation and
that new awropriate tests should be developed. Also monitoring of full
scale structures should be urgently undertaken. As a first step in this
 355

direction the damage occurring to geotextiles during the construction stage

should be determined.
3.2. Joints and Connections
Less obvious than durability but just as important to the long term
behaviour of geotextile reinforcements is the behaviour of their joints
(the 1 inkages between reinforcerrents) and connections (the linkages between
reinforcements and other components). Many different types of joints and
connections exist including sewing, stapling, bodkins and g lueing . The
short term behaviour of joints has been reported on by Murray et al (1986)
but no long term studies are known to have been undertaken either at
laboratory or full scale.
Often joints can be avoided by careful dimensioning of reinforcements but
connecticns are generally unavoidable. The various types of connections
employed should be identified and appropriate means of measuring their
properties in both the short and lcng term developed.

4. Surrrnary
The testing requirements for the geotextile materials used as soil
reinforcements have been identified but it has been shewn that there are
many areas where no suitable test methods presently exist. It would be
most useful if in time a list of the testing requirements for geotextile
reinforcements made together with the identity of both the available
suitable test methods and test methods that require to be developed.

REFEREOCES
1. MURRAY, R.T. and McGOWN, A.: The Selection of Testing Preeedures for
the Specification of Geotextiles, Pree. II Int. Conf. on Geotextiles,
Las Vegas, Vol. II, pp. 291-296, 1982.
2. MURRAY,.R.T. and McGCWN, A.: Geotextile Test Procedures: Background
and Sustained Load Testing. Application Guide 5, Transport and Road
Research Laboratories, Department of Transport, U. K. 12 pp. 1987.
3. MURRAY, R.T., McGCWN, A., ANDRAWES, K.Z. and SWAN, D.: Testing Joints
in Geotextiles and Geogrids, Pree. III Int. Conf. on Geotextiles,
Vienna, Vol. III, pp. 731-736, 1986.
4. R.I.L.E.M. Committee SM-47 (Synthetic Membranes): Report on Meeting at
Milano, Italy, pp. 10, 1985.
5. Session 2B: International Standards, Proc. II Int. Conf. on
Geotextiles, Las Vegas, Vol. II, pp. 291-333, 1982.
6. STRUD~, J.: Basic Principles Underlying the Swiss Guidelines for the
Use of Geotextiles, Proc. II Int. Conf. on Geotextiles, Las Vegas,
Vol. II, pp. 301-305, 1982.
 REVIEW OF SESSION

EVALUATION OF MATERIAL PROPERTIES

When dealing with reinforced soil structures, knowledge of the behaviour

of the different components of the structure is essential. This is gained
by testing the materials either separately or in combination, using stan-
dard test methods and specially developed testing techniques. Due to the
importance of this subj ect, it is not surpr~s ~ng that most o·f the
published work deals with various aspects of material testing. According-
ly, and perhaps ironically, designers are now faced with the problem of
choosing suitable tests from which the relevant design parameters are to
be determined. This session deals with the wider aspect of these problems
by examining materials evaluation which of course includes testing.

Leflaive started the session and stated that there are four purposes for
material evaluation. These are a) to check the consistency of information
in order to avoid mistakes in material selection, b) to enable the
engineers to refer to past experience, c) to provide data for calcula-
tions, and d) to enable the prediction of construction problems.

He concentrated on the problem of obtaining the relevant data for calcula-

tions (or dimensioning). The specific data sought depend to a large
extent on the design concept chosen. For example, in the "discrete con-
stituent material" approach it is the test data relating to the properties
of the individual materials together with the interface properties which
are needed and used in the analytical model. Another design concept is
the "equivalent composite material approach" which uses properties deter-
mined from tests performed on the mixture of soil and reinforcements.
Although t~e validity of this approach is doubtful for multi-layered sys-
tems, it may be applicable to situations where the polymer is randomly
mixed with the soil such as in "Texsol" and the "mesh element" system.
Design parameters may also be obtained from model and prototype testing of
reinforced structures and used in what is known as "system simulation"
approach.

Leflaive raised the questions of materials evaluation with respect to con-

struction damage, ease of laying, mechanical and chemical degradation,
permeability, and construction of joints and connections. He stressed the
need to develop test methods capable of assisting the engineers in assess-
ing these aspects. It was interesting to learn from Bush that this type
of evaluation is underway for geogrids. His brief presentation of the
results obtained so far emphasised the value of such work in both design
and construction.

McGown drew attention to the importance of testing the sub-soil. A com-

pressible sub-soil can greatly influence the internal strains in the
structure thus complicating the problem. He also pointed out the change

357

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 357-358.
© 1988 by Kluwer Academic Publishers.
358

which occurred in the facing elements and connections from the early
compressible aluminum facings to present day large concrete units. These
developments, he urged, should be taken into consideration when
identifying the test requirements.

With regard to soil testing, McGown pointed out the importance of lateral
(tensile) strains in reinforced structures and argued that it is therefore
essential to study the lateral strain behaviour of a range of typical
granular soils. He explained that the 10% criterion for tensile strains
in reinforced systems originated from the fact that at about 10% lateral
strain the angle of internal friction reaches its ~cv value. This
aspect of soil testing, particularly at low stresses, has received little
attention so far.

When dealing with soil reinforcement interaction there is a lack of agree-

ment whether the required design data should be obtained from ' pull out
tests or shear box tests. Scott expressed the opinion that both tests are
needed and presented details of a fully instrumented pullout test and a
shear box test on compacted clay and geogrid. Milligan provided some
valuable data regarding the interaction characteristics. He showed that
for sheet reinforcements the results obtained from direct shear test and
pullout test agree well, basically because the same interaction mechanism
operates in both tests. This is not the case for grid reinforcement where
the mechanisms are different due to the presence of opening in these
materials. However, it is accepted that there is a lack in the
understanding of the effects of stress level, soil type, compaction and
reinforcement stiffness on the interface characteristics.

It was interesting to observe that when McGown explained the various

methods for determining the stress-strain relationships of the
reinforcements, the discussion became lively. Basically, he explained in
detail the importance of testing wide width samples, the relevance of in-
isolation and in-soil tests and the effects of the temperature and the
strain rate on the measured parameters. He strongly recommended the use
of sustained load creep tests, both in-isolation and in-soil depending on
the soil and the geotextile types, for the determination of design data.
He highlighted the importance of the test methodology and of adopting a
valid technique for interpreting and processing the measurements. These
data are best presented to the eng ineers in the form of a load-strain
curve together with a time correction curve.

Leflaive argued in favour of testing wide width samples in-isolation

explaining that by adopting a width/length ratio of 5 and making correc-
tions to the measured strains, the results are similar to those obtained
using in-soil testing. McGown however, pointed out that although this may
be true in certain situations for tests conducted under constant rate of
strain, it is not the case for creep testing. In fact, from the
discussion which followed, it is apparent that the creep behaviour of
these materials is not clearly understood and that further work in this
area is urgently needed particularly concerning the effect of temperature.

This session has certainly confirmed the importance of material evaluation

and highlighted the difficulties involved. The most important outcome is
the identification of the areas where further research is needed in order
to achieve standardisation in approach and the areas which require long
term field observations.
EVALUATION OF MATERIAL PROPERTIES - DISCUSSION

DR. G.W.E. MILLIGAN,

University of Oxford, England.

In the analysis or design of a reinforced soil wall, the

properties of interaction bet\veen the soil and the reinforcement must be
taken into account, in addition to the properties of the soil and
reinforcement materials. Two simple limiting modes of interaction may
be identified (See Figure 1): -

(i) direct sliding, in which a block of reinforced

soil moves across a plane of reinforcement;

(ii) Pullout, in which a layer of reinforcement

pulls out of stationary soil within an anchor
zone.

Pullout test
.: .... ' .
'" " .
... .

Soil Reinforcement
I
Reinforcement

~
Shear test for direct sliding resistance

(a) sheet or strip reinforcement

(b) grid reinforcement

Fig. 1 Tests for soil-reinforcement interaction

359

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 359-361.
© 1988 by Kluwer Academic Publishers.
360

The tests normally performed to measure the interaction

properties are either modifications of the direct shear test, or pull-
out tests; it is important to be clear about the relevance of each.

The modified direct shear test will measure a coefficient of

direct sliding for any type of reinforcement. l,Jith a geotextile the
sliding is between soil and geotextile over the full plan area. With a
geogrid the sliding resistance is generated mostly by soil sliding over
soil in the apertures of the grid, and to a lesser extent by soil
sliding over the material of the grid. The test is relatively simple to
perform and interpret and is being adopted as a standard test in many
countries.

The pullout test models directly the second interaction mode,

but is more difficult to perform and interpret correctly. Special
apparatus is required, and results may be greatly influence'd by the
boundary conditions imposed in the test. With geotextiles the
interaction on either side of the reinforcement is very similar to that
in the direct shear test; bond coefficients obtained from correctly
performed and analysed pullout tests agree well with direct sliding
coefficients from modified direct shear tests. There is therefore no
need to perform pull-out tests for such materials; adequate data may be
obtained from the much simpler direct shear tests,

With grids the mode of interaction during pullout is quite

different from that in direct shear, as has been clearly established
from tests using a photoelastic technique (Dyer 1985) to visualise the
stress fields. In Figures 2 and 3 the light 'stripes' are aligned along
the directions of major (compressive) stress, and their brightness gives
a qualitative indication of the intensity of stress. A grid generates
bond in pullout by concentrations of bearing stress against the
transverse members of the grid, whereas direct sliding resistance is
generated by soil/soil and soil/reinforcement shear. Bond resistance of
grids can only be measured in pullout tests; with extensible
reinforcement in particular the interpretation of such tests requires
considerable care. For many design purposes it may be adequate to
predict the bond characteristics using the approach of Jewell et all
(1984 )'.

References

Dyer, M.R. (1985). Observations of the stress distribution in crushed

glass with applications to soil reinforcement. D. Phil. Thesis,
University of Oxford.

Jewell, R.A., Milligan, G.I-r.E., Sarsby, R,W. and Dubois, D. (1984)

Interaction between soils and geogrids. Symposium on Polymer
Grid Reinforcement in Civil Engineering. I.C.E. London.
 361

Fig_ 2 Stresses in pullout test using a grid

Fig- 3 Stresses in direct shear test using a grid

Analytical Techniques and Design Methods
 REINFORCED SOIL WALL ANALYSIS AND BEHAVIOUR

R.A. Jewell
University of Oxford

Abstract
The paper considers the equilibrium in reinforced soil walls.
Two solutions are presented which are expected to "bound" the
likely state of equilibrium in a wall. The magnitude and
distribution of the reinforcement forces and the soil stresses
are derived analytically, allowing the movement of the wall
face to be calculated. Two non dimensional charts are given
for the calculation of outward movement. The resulting
magnitude of movement predicted by the two equilibrium
solutions is close, so that the likely range of movement in a
wall is well defined by the analysis. Incremental wall
construction causes a "moving datum" for the outward movement
at the wall face due to reinforcement elongation. Thus,
incremental construction results in additional movement at the
wall face. This is the main distinction with propped walls
which do not suffer from incremental construction movement.
An analysis for the incremental movement is given. The
selection of compatible values of mobilised soil shearing
resistance and reinforcement force in a reinforced soil wall
is discussed, and new ideas put forward for the way the soil
is likely to behave in the reinforced zone. The application
of the concepts and the analyses introduced in the paper are
fully illustrated by the prediction for the Royal Military
College trial wall which is presented in a companion paper.

INTRODUCTION
1.1 Design and prediction
The detailed analysis for reinforced soil is complex - like
most soil mechanics - particularly when the soil is reinforced
by relatively extensible polymeric materials. The derivation
of design,methods is more straight forward; simple conserva-
tive assumptions can be made to allow for uncertain material
properties, boundary conditions, compatibility conditions, and
additional generous safety margins can be applied.

365

p, M, Jarrett and A, McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 365-408.
© 1988 by Kluwer Academic Publishers.
366

If a prediction of the equilibrium or serviceability state in

a reinforced soil structure is to be made then there is no
alternative but to try and evaluate the uncertainties.
Conservative design simplifications applied to prediction
would lead to overpredicted reinforcement forces and
deformations. This paper examines the prediction of equilib-
rium in reinforced soil walls.
1.2 Organisation of the paper
In the paper the behaviour of reinforced soil walls is
examined in several stages. The starting point is the concept
of equilibrium in soil and in reinforced soil. This allows
possible sets of forces required to maintain equilibrium in
soil with a constant mobilised shearing resistance to be
derived. These forces would be provided by the reinforcement.
Two possible equilibrium states for reinforced soil walls are
derived. The magnitude and distribution of the reinforcement
forces and the stresses in the soil are given. The two
solutions represent two extreme cases for equilibrium in
reinforced soil which, it is suggested, are likely to "bound"
the actual equilibrium.
The next stage is to examine the magnitude of outward movement
at the face of the wall which would be implied by the required
reinforcement force distribution and the reinforcement
stiffness. The outward movement can be determined analytical-
ly and the results for the two equilibrium solutions are
presented in non dimensional charts. The deformation in the
reinforced zone is assumed to be controlled by the
reinforcement elongation; the soil shearing resistance is
assumed constant and independent of the deformation.
The influence of incremental layered construction is examined
next, particularly to see what datum for movement might be
used (even for movement due only to reinforcement elongation)
when the incremental facings are aligned over built facings
which hpve already undergone movement.
To determine the likely equilibrium combination of the
mobilised soil shearing resistance and reinforcement force in
a reinforced soil wall, the relationship between the
deformation and the shearing resistance in soil has to be
examined. The link with the mobilised reinforcement force is
through compatible strain. At best the reinforcement strain
would equal the tensile strain occurring in the soil in the
direction of the reinforcement. The overall equilibrium in a
reinforced soil wall can be represented on a compatibility
curve.
These ideas are fully illustrated by their application to the
prediction of the behaviour of the incremental trial wall
built at the Royal Military College, Kingston. The analysis
is presented in a companion paper "Analysis and predicted
behaviour for the Royal Military College trial wall".
 367

EQUILIBRIUM STRESS DISTRIBUTIONS

2.1 Equilibrium in soil

Soil is a frictional material. An element of soil will remain
in static equilibrium if the stress ratio remains below the
available frictional shearing resistance. Frictional shearing
resistance depends on the type of loading applied to the soil,
and plane strain is the relevant loading condition for walls.
Only frictional soils are considered in this paper.
Frictional shearing resistance depends on effective stresses
in the soil. For simplicity, the dry case is examined in this
paper so that all the stresses are effective stresses.
The Mohr circle of stress represents the state of stress
equilibrium in a soil element. The case of soil in
equilibrium with a mobilised frictional resistance ~ less than
the plane strain frictional resistance rfJ p , is shown in Fig. 1.
1:

C1

Fig. 1. Equilibrium stresses in frictional soil

The maximum stress ratio in the soil is often expressed as a

ratio of the shear stress and normal stress acting on the
plane of maximum stress obliquity, Fig. 1., giving
tp
-= tan¢ (1 )
C1 p

but the stress ratio in the soil can also be expressed more
generally in terms of the principal stresses to give

(2)
368

It is usual to define the factor of safety for the soil in

terms of the ratio of the shearing stresses on the plane of
maximum stress obliquity compared with the maximum available
shearing resistance ~P' , so that, for the plane strain case,
tan ¢ ps
Factor of safety = ----'- (3)
tan¢

2.2 Equilibrium in reinforced soil

Reinforcement modifies the stresses in soil. A simple
illustration is a direct shear test on an element of
reinforced SOil, as described by Jewell and Wroth (1987), Fig.
2. The reinforcement force P r acting across the plane of
maximum stress obliquity in the soil reduces the shear force
that the soil must support and increases the no~al force in
the soil.

(a)
Th~ effect of reinforcement is twofold:

(1) To reduce the str~sses causing failure

PRsinS
1: = 1: yx - ---,;:s

(2) To increase the stresses resisting failure

a= a + PRcosS
YY As
(b)

Fig. 2. Reinforced soil in direct shear

The overall equilibrium in the soil should be expressed in

terms of the forces acting across the plane in the soil being
examined so as not to presume a uniform state of stress.
Uniform stress is unlikely in reinforced soil.
However, it is a convenient concept to consider the influence
of the reinforcement in terms of additional stresses in the
soil. In the shear box, on a plan area of soil As, the
addi tional resisting shear stress provided directly by the
reinforcement is
 369

(4)

and the additional normal stress provided by the reinforcement

is
P r case
(J =--- (5)
R As

These additional stress resultants acting in the soil enable

the soil to remain in equilibrium under increased shear
loading.
2.3 Equilibrium in a vertical reinforced soil wall
Previous concepts
The results of limit equilibrium analysis indicate three
important zones of stress in the soil of a reinforced slope or
wall, Fig. 3a.
Locus of zero required force

Critical surface
has greatest
required force

Similar pattern
} of force to
above
competent foundation

(a) Zone: (b)

1. High required forces \111 Magnitude of force

2. Decreasing required forces _._ Most critical mechanism
3. No required forces _______ Locus of zero force
Fig. 3. (a) Zones of required force for equilibrium and (b)
resulting reinforcement force distribution in a wall
(Jewell,1985)

There are two surfaces through the soil that separate the
three zones. Firstly, deep in the soil is a surface beyond
which no additional stresses are required from the
reinforcement to maintain equilibrium. This is a locus of
zero required force. Secondly, through the toe of the slope
there is a surface which requires the greatest total
reinforcement force to maintain equilibrium. This is the most
370

critical surface through the toe, because the greatest

reinforcement force is required to maintain equilibrium on
this surface.

Reinforcement spacing; calculation of force in reinforcement

layers
In fact, there are a whole family of geometrically similar
surfaces parallel to the most critical surface intersecting
the slope face above the toe.
For the vertical case the surfaces are plane and at an angle
B= (4S+~/2) to the horizontal, Fig. 4a. The similarity of
these surfaces through the soil implies that the reinforcement
must be providing horizontal stress in the soil which
increases linearly with depth, Fig. 4b.

Uniform surcharge
{ CT h } required
o ,-.-----r-.......-
I
I
""----
I Available stress
I with uniform
I spacing
I
I
I

,"

Fig. 4. (a) Most critical mechanism and similar surfaces

through the face. (b) Triangular distribution of required
stress from similarity. (c) Balanced distribution of
reinforcement satisfying required stresses both locally and
overall.
 371

The stress required for equilibrium in the soil is provided by

the force in the reinforcement layers. Jewell et al (1984)
introduced the idea of a balanced distribution of reinforce-
ment; the idea is that the reinforcement should be spaced to
satisfy both local and overall equilibrium.
Locally each reinforcement layer can only provide the required
soil stresses for one half the vertical spacing above and
below the reinforcement layer. To make the most of
reinforcement layers with a fixed allowable force the spacing
must be reduced with depth so that the full allowable force
can be mobilised in each layer, Fig. 4c. This would give an
ideal spacing arrangement because each reinforcement layer is
able to mobilise its full allowable force.
If uniform spacing was chosen for the reinforcement the
required spacing would be set by the deepest reinforcement
layer at the bottom of the wall. Half way up the wall the
required reinforcement force would have dropped by half, Fig.
4b, and the reinforcement would have twice the required
capacity. It is convenient to use a diagram of required soil
stresses and available soil stress from the reinforcement,
similar to Fig. 4b, to assess the chosen spacing arrangement.
Conversly, this diagram can be used to determine the mobilised
force in each layer in a chosen spacing arrangement.
In a balanced design the available stress provided by the
reinforcement should always be greater than the required
stress for local equilibrium at every depth in the wall, to
avoid local overstressing of the reinforcement. Local rupture
of a reinforcement can result in overall collapse, section
3.2.
In Fig. 3b the reinforcement extends back to the locus of zero
required force. This reinforcement layout can be thought of
as having ideal length because the reinforcement is provided
everywhere in the soil that reinforcement force is required
to locally maintain equilibrium. The case of equilibrium with
reinforce~ent of ideal length is considered below.

2.4 Equilibrium with ideal length reinforcement

The starting point is soil with an assumed constant mobilised
angle of friction and built to a vertical face. The aim is to
derive a state of equilibrium for the soil everywhere and
thereby to determine the required reinforcement layout.
The two bounding surfaces in the soil (see Fig. 3a) are both
plane surfaces in this case. The locus of zero required force
is a plane through the toe inclined to the horizontal at the
angle of friction, plane DE in Fig. Sa. The most critical
surface through the toe is the critical Coulomb wedge, a plane
inclined' at an angle e = 45 + ~/2 to the horizontal, plane DB in
Fig. Sa.
372

A B c o

a..

A B A C E

~.P ~RO
It,::IZ Q

o o 0
b. c. d..

Fig. 5. Bounding surfaces for reinforcement of ideal length

The simplest way to consider the state of equilibrium is in

terms of the gross required (reinforcement) force for
equilibrium. A reinforcement force equal to the active force
Pa is needed to hold the zone OAB in equilibrium on the plane
OB, Fig. 5b. A smaller reinforcement force is needed to hold
the zone OAC in equilibrium on the plane OC, Fig. 5c.
Progressively smaller force is needed for equilibrium on
flatter planes 00, until no reinforcement force at all is
required for equilibrium of the block OAE on the plane OE,
Fig. 5d.
The variation of the overall required reinforcement force
depend~ on the angle a to the horizontal of the plane surface
in the soil through the toe, Fig. 6.
From a wedge analysis the overall required force for
equilibrium is
(6)

which can be expressed as

(p ) = yH 2 tan(e- ~)
(7)
R e 2 tan e

Equa~ions (6) and (7) describe the variation of the required

reinforcement force between the most critical plane through
the toe and the locus of zero required force. The required
reinforcement force is constant between the most critical
plane and the face. The state of equilibrium, the
 373

reinforcement layout and the variation of force along the

reinforcement is illustrated in Fig. 7 for reinforcement with
ideal length.

1
H

t
R

Fig. 6. Wedge analysis for the required reinforcement force

The ideal length state of equilibrium is the analytical

counterpart to the pattern of reinforcement force in a
reinforced soil wall deduced from limit equilibrium analysis
and described by Jewell (1985), Fig. 3.
Equation (7) gives the variation of the total required
reinforcement force with the angle e through the toe. The
expression for the force P, in a single layer of
reinforcement, which supports a force P at the most critical
plane, is

which is valid in the range (45+\D/2)~e~\D. Remember that P,= P

when e> 45+16/2.
Implied stress distribution for ideal length layout
The state of stress in the reinforced soil can also be
identified. The Mohr circle for stress for soil elements
along a line through the toe at an angle e to the horizontal,
which is a plane of maximum stress obliquity, is shown in Fig.
8a. From the pole for planes , it is pos sible to deduce the
orientation of the second plane of maximum stress obliquity in
the soil and determine the orientation of the principal
stresses' in the soil. The resulting stress distribution is
illustrated in Fig. 8b by the pattern of planes of maximum
stress obliquity and the principal stress directions.
374

Most critical plane

Locus of zero
required force

eqn.(8)

Pr/p

Position on reinforcement
Fig. 7. Reinforcement layout and force distribution for the
ideal length case.

2.5 Equilibrium with truncated length reinforcement

Consequences of truncation
Truncation of the reinforcement short of the locus of zero
requir~d force changes the pattern of equilibrium in the
reinforced soil. For example, consider the reinforcement
layout for ideal length illustrated in Fig. 7, but with the
maximum reinforcement length limited to Lmax. In the zone
where the reinforcement has been omitted, the soil equilibrium
will be equivalent to the active state behind a retaining
wall, Fig. 9. The soil will exert active pressure on the back
of the reinforced zone.
Fig. 9 illustrates the two consequences of truncation of the
reinforcement length. Firstly, when reinforcement force is
omitted higher up on a surface through the reinforced soil the
lower reinforcement layers have to provide the additional
force. - the required force is shed to lower layers. Secondly,
the active thrust from the unreinforced soil resting on the
back of the reinforced zone changes the position of the locus
of zero required force - it extends more deeply into the soil.
 375

(o) (b)

Fig. 8. Mohr circle for stress, and planes of maximum stress

obliquity in the soil for the ideal length reinforcement case.

In spite of the imbalances which result from truncating the

reinforcement length, it is likely that the truncated
reinforcement layout will give a more efficient use of the
reinforcement layers as explained below.

Truncation without increasing the required force

In the ideal length arrangement the most critical mechanism
through the toe governs the gross required reinforcement
force. Each reinforcement layer has an allowable force and
spacing selected to hold the most critical mechanism in
equilibrium. This magnitude of force is available along the
whole length of the reinforcement where bond with the soil
permi ts . Over much of the reinforcement length in the ideal
length layout the mobilised reinforcement force is less than
the available force.
Truncating the length of upper reinforcement layers sheds
required ' force to lower reinforcement layers and requires
addi tional force to be carried by the lower layers; but as
long as the additional force does not increase the required
force for any single layer above the allowable value then the
reinforcement already has sufficient capacity to carry the
additional load.
376

-Lmax-

1 Additional soil
H
loading
ll.PR - Extra required
force

Fig. 9. Illustration of the consequence of truncating the

reinforcement length.

In other words reinforcement length can be truncated higher in

the wall without need for increase in the capacity of the
reinforcement until the resulting load shedding brings the
required force on lower reinforcement layers to the allowable
force selected to support the most critical mechanism through
the toe. Thus reinforcement length can be saved higher in the
wall wi thout requiring a change in the reinforcement
properties or cross section.
It must be emphasised, however, that some additional
reinforcement length is also required lower in the wall as a
result of truncation.
Truncated length layout
To illustrate these ideas, consider the limiting case of
reinforcement layers that can mobilise the allowable
reinforcement force over a negligible bond length - so that
the available reinforcement force is constant over the full
length ~f the reinforement. Also, for simplicity, assume that
the active state in the unreinforced soil behind the
reinforced zone is equivalent to soil behind a smooth
retaining wall (ie. with vertical and horizontal principal
stresses).
The equilibrium in the zone between the front of the wall and
the most critical mechanism is the same as before and the same
reinforcement spacing and material properties would be chosen
as for the layout for reinforcement of ideal length, zone OAB
Fig. lOa.
With truncated reinforcement there will be two sections on any
plane,through the toe at an angle e«45+¢/2) to the horizontal,
Fig. lOb. On the portion OC the reinforcement will provide
the full allowable force on the reinforcement layers
intersected. On the portion CC" the soil is unreinforced and
held in equilibrium by exerting active stresses on the back of
the reinforced block, illustrated by the vertical plane CB'.
 377

A S s' e"
''-~

1
H
(PR) 08

,
/. ....

(e) (b)

(P R }c's

(c)
Fig. 10. Equilibrium with truncated reinforcement layers
supporting the full active force and with negligible bond
length

The position of C on a line at an angle e can be found so that

there is force equilibrium between the soil self weight
forces, the active driving force of the unreinforced soil and
the reinforcement layers acting with their full allowable
force.
The equilibrium state for the soil segment OBB'C is
illustrated in Fig. lOco The segment has two net outward
disturbing forces to support: firstly, the net reinforcement
force which acts on OB but not on OC; secondly, the outward
active earth force from the unreinforced soil acting on CB'.
The soil segment OBB'C can remain in equilibrium because the
angle of the plane OC is less than OB, so that greater
shearing resistance can be generated on OC by the same
horizontal force. The greater net shearing resistance along
OC, in combination with the additional self weight of the soil
segment «L1W), enables the additional outward thrusts on the
soil to be held in equilibrium.
Repetition of the calculation on flatter soil segments, and
with the simple assumption of a net active thrust for the
smooth case from the unreinforced fill, provides a
distribution of the reinforcement length with depth in the
wall required to maintain force equilibrium. The results in
Fig. lla. are for rp= 30·.
378

A 8 A 8

o E o E
(a) ( b)
Fig. 11. (a) Required reinforcement length for truncated
layers with negligible bond length. (b) Equivalent stress
field in the reinforced soil.

The required reinforcement length to provide force equilibrium

can be expressed analytically. The derivation is not repeated
here. One way to express the result is with a pair of
equations giving the required reinforcement length (LlH) at a
depth below the crest of the wall (zl H) in terms of a plane
through the roe at an angle e to the horizontal. The
expression for the required reinforcement length is
LIz ( 9)
-=-(1--)
H tane H

which applies at a depth

z tan etan 2 ( 4S - ¢/2) - tan(e- ¢)
(10)
H 2tan etan 2( 4S - ¢/2) - tan(e - ¢)

Implied stress distribution for truncated length reinforcement

As explained for the ideal length case, each line segment
through the toe is a plane of maximum stress obliquity and so
the orientation of the principal axes of stress in the
reinforced soil can be determined. The planes of maximum
stress obliquity in the reinforced soil are drawn in Fig. llb,
where the directions of principal stress are indicated.
The simple assumption made about the unreinforced fill behind
the reinforced zone results in vertical and horizontal
principal stresses in the unreinforced soil. Because this was
an assumption there is a lack of fit between the two zones of
stress along the plane BCDE, Fig. lIb.
 379

The implied stress distribution for the reinforced soil shows

that the wall resists the outward thrust from the unreinforced
soil by taking advantage of the additional shearing resistance
available from the reinforcement forces acting on less steeply
inclined planes through the wall toe.
The required reinforcement length distribution in Fig. 11a is
shown for soil with a mobilised angle of friction 30'. The
required reinforcement length in the truncated length case for
soil friction angles in the range 20' to 50' are given in Fig.
12.
TRUNCATED REINFORCEMENT LENGTH

3; 3~
0

~~\
,..... 50 45 40
I
..........
N
~ 0.2
I-
(f)
I
W
0:::
u 0.4 / I / I \
-'
-' // / / I
~
0.6 J J / J / I
/ II I I
:l:
g
w / I
CD
I 0.8 II // I / V /
/ // V / / I
l-
n.
w

~ ~V /
Cl
/
./. ./
o 0.2 0.4 0.6 0.8

REINFORCEMENT LENGTH (L/H)

Fig. 12. Required minimum reinforcement length for force
equilibrium (infinite bond case).

The reinforcement lengths shown in Fig. 12 are an absolute

mlnlmum for the appropriate soil friction angles and
reinforcement designed to carry active stresses. The
important omission is that they only represent force
equilibrium and not moment equilibrium. The purpose of these
results is to illustrate ways that reinforcement might
maintain equilibrium in reinforced soil. This is useful for
assessing possible reinforcement layouts.
Equilibrium could be maintained with shorter reinforcement
lengths than those given in Fig. 12 for the truncated length
case if the allowable reinforcement force was higher than that
needed to support the active soil stresses. In this case
there would be load shedding to lower reinforcement layers
even on the most critical plane.
380

2.6 Stress transfer between reinforcement and soil in a wall

The stress distributions in Figs. 8 and 11 are for the
reinforced soil, that is they represent the stresses in the
soil in equilibrium with the mobilised reinforcement forces.
The reinforcement is (conceptually) distributed through the
soil rather than in discrete layers. However it is still
possible to identify the overall pattern of stress transfer
from the reinforcement to the soil.
The central feature of both the derived equilibrium states is
the increasing horizontal outward shear stress in the soil
behind the most critical plane and this causes the principal
stress directions to rotate.
In the uniform zone between the wall face and the most
critical plane the reinforcement force is constant and there
is no transfer of stress between the soil and the
reinforcement. The transfer happens at the wall face where
the required horizontal stress is applied to the soil by the
wall face, which in turn is supported by the reinforcement,
Fig. 13a. This is an ideal facing boundary.

~-
/3
-
SoH

'i--
Facing ~

t
--~3
4Reintorcement

Soil
:;-- Soil PRI - PR2
lrt: R =
6L
(0) ( b)

(e)

Fig. 13. Illustration of the load transfer mechanism between

the reinforcement and the soil.
 381

The reinforcement axial force is distributed back to the soil

behind the most critical plane. The transfer of load occurs
by shear stress developing between the reinforcement and the
soil as illustrated in Fig. 13b, drawn for the case of wide
width or sheet reinforcement. Overall, the axial force from
the reinforcement layers are transmitted down into the soil.
This is illustrated in Fig. 13c for the ideal length case.
At any level in the soil behind the most critical plane the
principal stress directions in the reinforced soil are
inclined from the vertical. A layer of wide width
reinforcement would experience outward shear stress from the
soil above which would be transmitted through the reinforce-
ment to the soil below, see the detail in Fig. 13c. The
reinforcement in turn also transmits some axial force to the
soil by increasing the outward shear stress on the soil below.
The soil below the reinforcement layer is able to support the
additional outward shear stress. Firstly, the principal
stresses rotate further with depth in the soil which increases
the ratio of shear stress to vertical stress. Secondly, the
ability of the soil to carry shear stress increases with depth
as the vertical self-weight soil stresses increase.

COLLAPSE LIMITS

3.1 Collapse without reinforcement rupture

Collapse will occur in a reinforced soil wall if on a
potential failure mechanism there are insufficient available
resisting forces. This may arise in two ways.
Firstly, if there is insufficient reinforcement length the
bond can limit the mobilised reinforcement forces to below the
value required to maintain equilibrium. Lack of equilibrium
would leati to continued shear in the soil and collapse.
Secondly, the shearing resistance of soil across the surface
of a layer of reinforcement in the ground can be less than the
shearing resistance of the soil alone. In this case the layer
of reinforcement provides a plane of reduced shearing
resistance and if this does not balance the outward thrust of
the unreinforced soil an outward direct sliding shear failure
would develop.
3.2 Collapse with reinforcement rupture
Examples of equilibrium stress states in reinforced soil have
been illustrated in Figs. 8 and 11. This pattern of stress
equilibrium is likely to exist in a reinforced soil wall
immediately prior to collapse caused by reinforcement rupture.
If a reinforcement layer were to rupture (due to a local
weakness, say) the local loss of equilibrium would cause the
soil immediately above and below the rupture point to be
overstressed and out of equilibrium. Local strains would
382

develop as the soil locally redistributed the out of balance

stresses. This mayor may not cause rupture in an adjacent
reinforcement layer before equilibrium is established once
more.
Thus collapse caused by reinforcement rupture is started by
local loss of equilibrium. Collapse caused by bond and
outward direct sliding is started by general loss of
equilibrium.
Collapse by reinforcement rupture propogates progressively
from the local source of instability.
This is the reason why reinforcement layers should be
distributed through the soil in a balanced way to everywhere
satisfy local as well as overall equilibrium, (section 2.3).

WALL DEFORMATION

4.1 Sources of movement

Deformation of a reinforced soil wall should be expected from
the three sources illustrated in Fig. 14.

Reinforced soil
Wi th extensible polymer reinforcement materials the greatest
deformation is likely to develop in the reinforced soil, zones
A and A' Fig. 14. Wi th reinforcement attached to the facing
panels, the elongation in the reinforcement layers should be a
good measure of the lateral deformation in the reinforced
zone.
Between the most critical plane and the wall face, zone A in
Fig. 14, the principal stresses in the reinforced soil are
approximately vertical and horizontal. This zone is likely to
deform • relatively uniformly causing both horizontal and
vertical movement at the face, Fig. 14b.
Behind the most critical mechanism, zone A' in Fig. 14, the
principal stresses in the reinforced soil are inclined from
the vertical, very markedly towards the base of the wall. The
vertical movement in this zone will be much reduced compared
with the horizontal movement.

Backfill and foundation

With truncated reinforcement the soil behind the reinforced
zone must come into an active state of equilibrium, zone B in
Fig. l4. This requires strain in the soil which will push the
reinforced zone outwards.
Finally, the reinforced zone, and the unreinforced soil
behind, rest on the foundation, zone C in Fig. 14. Movement
caused by overstressing the foundation beneath the reinforced
zone, or due to overall settlement in the foundation will
cause additional "rigid body" movements at the wall face.
 383

(a) (b)

Fig. 14. Zones causing deformation at the face of a

reinforced soil wall.

Care is needed with field observations of face movements in

reinforced soil walls to separate the movement due to
deformation of the reinforced soil, deformation behind the
reinforced zone and the movement resulting from foundation
deformation.
4.2 Deformation predictions for the reinforced zone
The two equilibrium stress fields for reinforced soil walls
with ideal length and truncated length both directly give the
magnitude and distribution of reinforcement force in the
reinforcement layers. If the stiffness of the reinforcement
is known, the strain along the reinforcement may be calculated
and hence the elongation in the reinforcement layer can be
derived analytically. With no relative slippage between the
reinforcement layer and the soil, the elongation in the
reinforcement would equal the horizontal component of movement
at the wall face caused by deformation in the reinforced soil.
It is convenient to represent the horizontal outward movement
6 as a proportion of the wall height H. The reinforcement
stiffness K and axial force p govern the maximum reinforcement
strain. The horizontal outward movement of the wall face due
to deformation in the reinforced zone can be expressed by the
non dimensional parameter
384

The appropriate units for the load in geotextile and grid

reinforcement are force per unit width (kN/m) • The
reinforcement strain should be expressed as a number (0.01)
when evaluating parameters such as stiffness, rather than as a
percentage (1%).
The reinforcement spacing arrangement influences the mobilised
force in the reinforcement layers. With ideal spacing the
reinforcement spacing is chosen so that the full allowable
reinforcement force is mobilised in every layer. With uniform
spacing the full allowable reinforcement force is only
mobilised in the lowest reinforcement layer; progressively
less force is required in higher layers in the wall to
maintain equilibrium, which results in progressively less
outward movement.
The non dimensional charts for outward movement presented
below can be used for any chosen reinforcement spacing by
using the mobilised reinforcement force P mob calculated for
local equilibrium in each layer, section .

IDEAL REINFORCEMENT LENGTH CASE

I
~ 0.2t-~---r-r~70~~--yL-.~;'--r-~
E
C>:
u 0.4 t-~---t-.HiT-7I--;1'<f--rt----+---+--l--~

~m 0.6 t-~-H7'lt'-r¥--A'-----+-1----+---+---I--1

0.8 t--UAW7f---+--I--+--+-----+---+--I-~

1 ~
o I 0:4 I I I 0:8 I I I :£"
1 1.6 2

OUTWARD MOVEMENT ( ~~

Fig. 15. Non dimensional outward movement at the face due to

deformation in the reinforced zone. Reinforcement with ideal
length.

4.3 Iq.eal length

The non dimensional chart for outward movement for
reinforcement with ideal length is shown in Fig. 15. The
variation of outward movement with depth in the wall closely
reflects the variation of reinforcement length with depth,
 385

Fig. 7. The results are for a range of angles of friction for

the soil and, . as would be expected, the soil with lower
shearing resistance requires a greater zone of soil to be
reinforced which results in greater deformation in the
reinforced zone.

4.4 Truncated reinforcement length

The non dimensional chart for outward movement in the case of
reinforcement with truncated length is given in Fig. 16.
Again, the variation of outward movement with depth closely
reflects the variation of the reinforcement length with depth,
Fig . 9.

TRUNCATED REINFORCEMENT LENGTH CASE

0
\
........
:::r:
50 45 40 35
3~2~2~
'-...
N
'-' 0.2 , \
f- I
~
a::
0.4 1I / J \
// /
(,)
II
...J
...J /
-'- / V / / /
~
:3: 0.6
g
w / VI I I / I
II V/ I
m
:::r: V /
0.8
L /~ / / / I
f-
a..
W
Cl

L 0 Y/ / /
V
o 0.2 0.4 0.6 0.8

OUTWARD MOVEMENT (~~

Fig . 16. Non dimensional outward movement at the face due to
deformation in the reinforced zone. Reinforcement with
truncated length.

The two charts in Figs. 15 and 16 can be used to estimate the

outward movement caused by elongation of the reinforcement to
maintain equilibrium. The movement depends only on the
mobilised soil shearing resistance ~ and the mobilised
reinforcement tensile strain PIK. The distribution . of outward
movement with depth below the wall face should be calculated
from both charts. The expected movement should be within the
range calculated.
386

4.5 Example application

To illustrate the use of the non dimensional charts, consider
an 8 metre high wall, with no surcharge, soil unit weight 19
kN/m3 and designed for a mobilised angle of friction 30'. The
reinforcement design allowable load is 26 kN/m and the
stiffness is 1000 kN/m (representing material properties in
the ground, at the end of the design life, at the design
temperature). This stiffness gives an strain 0.026 (ie 2.6%)
when the allowable load is mobilised in the reinforcement.
The gross required force for equilibrium on the most critical
plane is
K H2
( P R) rgquirgd aY =202 kNlm
2

which can be provided by 8 reinforcement layers if ideal

spacing is used.
The minimum required spacing at the base of the wall for a
unifor.m spacing layout is given by the expression (for a wall
without surcharge)

which for the example above gives Sv = 0.53 m, indicating that

15 layers of reinforcement are required.
If the truncated length arrangement is used, the face movement
for the two spacing arrangements can be calculated from Fig.
16. The curve for the outward movement is the same as in the
Fig. 16 for the ideal spacing case, except the actual value
for PIK for the wall should be used. With unifor.m spacing the
mobilised reinforcement force P decreases as a proportion of
the depth below the wall crest, directly reducing PIK and
hence the outward movement.
The results for the example are shown in Fig. 17. Two scales
for the outward movement are indicated
(1) outward movement as a function of wall height
(2) the magnitude of the outward movement in mm.
The first scale is probably the most general representation of
outward movement and for the design cases in Fig. 17 the
maximum outward movement is approximately 1.6% and 1% of the
wall height for the designs with 8 and 15 reinforcement layers
respectively.
 387

0
.01 .02 (~

e I'"
I 1\

'"
0.2
I-
UJ
W
u
c::
0.4 ~
::l
~
\ ~
:;= 0.6 1\ !
9
w
CD

~w O.B
\ / •
+
IDEAL SPACING
UNIFORM SPACING

I..t-V
0

o 40 80 120 160 200

FACE MOVEMENT (mm)

Fig. 17. Illustrative results for an 8 metre high wall.

Reinforcement layout with truncated length~ movement calculat-
ed for ideal spacing and uniform spacing arrangements.

4.6 Use with uniform surcharge loading

The non dimensional movement charts may be applied to walls
subject to uniform surcharge load. ,All that is required is
for the mobilised reinforcement force in each layer to be
calculated from a local equilibrium balance allowing for the
surcharge load (section 2.3).
The procedure is illustrated in detail in the companion paper
describin9 the predictions for the RMC trial wall.

INFLUENCE OF THE CONSTRUCTION SEQUENCE

5.1 Introduction
The analysis for reinforced soil wall behaviour has been
examined in stages. Firstly, equilibrium in a reinforced soil
wall has been examined for soil with a fixed and constant
angle of friction. Secondly, the state of equilibrium in the
"constructed" wall has been used to deduce the magnitude and
distribution of the reinforcement forces and hence the
elongation in the reinforcement layers.
388

At the very least, the lateral deformation in the reinforced

soil must equal the reinforcement elongations, and the
distribution of reinforcement layer elongations were plotted
in Figs. 15 and 16 to represent the outward movements at the
wall face caused by deformation in the reinforced zone.
However, walls are not "created" but are built, and in this
section the affect of incremental construction on outward
movements will be discussed.
5.2 Incremental construction
The self-weight loading experienced by the soil during
incremental construction progresses in a relatively uniform
way as illustrated in Fig. 18. The position of the most
critical mechanism, for example, stays constant in the soil
throughout construction. The vertical stresses in the soil
increase with construction and so must the horizontal stresses
to provide equilibrium. Thus the reinforcement force and
elongation builds up throughout construction.

IT
t It-------+--~
z
Most critical
plane

------£:
--- --- a
Fill
construction

Fig. ' 18 Progressive increase in stress during incremental

construction.

When it comes to examining outward movement at the face due to

deformation in the reinforced zone, it is apparent that there
is a moving datum that must be taken into account. This
occurs because after constructing the first incremental lift
of a wall, which deforms and moves outward, the subsequent
facing is aligned over the already deformed lower facing.
For each additional fill layer, it is the outward movement
developed in the layer immediately below, over which the
facing is aligned, that causes the move in the datum.
 389

In principle, a cumulative adjustment should be made at each

incremental level to allow for the slightly moved starting
point of each unstressed layer of reinforcement. This acts to
increase the outward face movements higher in the wall, as
will be illustrated.
Any calculation for this effect can only be approximate
because of the many factors which can significantly alter the
magni tude of the deformation in each increment of construc-
tion. For example,
the compaction loading is significant but this is
likely to be quite variable from layer to layer
the connection between the facings is likely to
significantly alter the amount of movement and the
mode of movement - from translation to rotation
if the panel rotates rather than translates the
movement at the top of the panel could change· by a
factor of at least two - and the alignment for the
next facing panel is to the top of the previous
panel.
Approximate calculation for incremental construction dis-
placement
An approximate calculation for the average incremental
movement caused by the construction of each layer is given
below. Depending on whether the panel rotates or translates
the movement at the top of the panel may be a factor of two
larger. Thus only an approximate assessment is possible for
what will actually happen in the field.
The calculation examines each increment of construction as if
it were a small, independent reinforced soil wall. It is the
deformation in the soil zone defined by the panel toe that
causes the moving datum.
In the calculation, the reinforcement force required to
maintain 'equilibrium for the panel under the soil weight and
the compaction surcharge loading is estimated. The length of
reinforcement over which the force will act ~s calculated
based on the equilibrium stress states described earlier.
The incremental reinforcement force PINe required to maintain
equilibrium in a layer of construction of height H lNc is given
by

(11 )

where q, is a uniform surcharge representing the load from the

compaction, and n is the number of reinforcement layers
connected to the panel.
The reinforcement force acts over a length of reinforcement
given by the equation
390

_( Qs)tan(90-¢)
L 1NC - HINC+-y 2 (12)

based on the ideal length equilibrium state.

The incremental displacement or outward movement caused by the
construction of each layer is
P INcLINC
incremental displacement = K (13)

where K is the stiffness of the reinforcement, taking account

of the rather short loading period for an increment of
construction compared with the long term design life for the
wall.
The above analysis is for a translating panel; the incremental
displacement at the top of the panel given by equation (13)
should be doubled if the panel rotates around the base rather
than translates.
Example
For the 8 metre high wall described earlier, and for the
uniform spacing reinforcement layout, consider incremental
construction in 8 layers ( ie H INC = 1 m . There are two
reinforcement layers in each increment n = 2). Use the same
mobilised soil shearing resistance and reinforcement stiffness
as before and assume that the compaction is equivalent to a
uniform surcharge load q, = 5 kNlm2.
The result when these values are substituted into equations
(11) and (12) are
P INC = 2.4kNlm

allowing for two layers of reinforcement per panel, and

LINC = 1.1 m

For reinforcement with a stiffness K= lOOOkNlm the incremental

displacement is 2.6 mm from equation (13), or 5.2 mm if panel
rotation rather than translation occurs.
The overall outward movement due to deformation in the
reinforced zone for uniform spacing for the truncated length
arrangement can now be illustrated, Fig. 19. The cumulative
addi tional outward movement due to incremental construction
amounts to 18 mm or 36 mm depending on the fixing between
panels.
The effect of construction on the outward movement is much
more Significant in the upper part of the wall.
 391

The incremental displacement depends on the number of

reinforcement layers in each layer of construction. For
uniform spacing this is the same up the wall and the
incremental displacement for each layer is the same. For
other spacing arrangements, such as ideal spacing, the number
of reinforcement layers reduces higher in the wall, requiring
calculation of the incremental displacement for each layer.
Because with ideal spacing the reinforcement is more widely
spaced higher in the wall the incremental displacement will
also be greater higher in the wall. This further increases
the affect of incremental construction on the outward movement
towards the top of the wall.
0

..--...
:::c
"-
~ ~ l\

"0l\
N 0.2
'-"

tilw
0:::
0.4
\ ~\
()
..J
..J
«
.3:
3: 0.6 ~
W
0
..J
III
- \', +
• NO INCREMENTAL
DISPLACEMENT
:::c 0.8
TRANSLATIONAL
DISPLACEMENT
I-
a...
~/ <>

)
W ROTATIONAL
CI DISPLACEMENT

o 40 80 120 160 200

FACE MOVEMENT (mm)

Fig. 19. The effect of incremental displacement due to
incremental construction on the outward movement of the face
due to deformation of the reinforced zone.
392

SAND STRENGTH AND DEFORMATION

6.1 Introduction
There are almost any number of distributions of mobilised soil
frictional resistance and mobilised reinforcement force which
will provide equilibrium in a reinforced soil wall. For
analysis, the problem has to be simplified by assuming that
the mobilised frictional resistance is constant (at least in
specified zones), and this technique has been used to
investigate reinforcement force distributions which would
provide equilibrium.
The two components providing equilibrium in the reinforced
soil, the soil and the reinforcement, are not independent of
one another, however, but linked through strain compatibility.
The mobilised frictional resistance in the soil depends
importantly on the strain which develops in the soil.
Likewise, the mobilised reinforcement force depends on the
strain in the reinforcement. If the unstressed reinforcement
is placed in soil, then the maximum possible strain in the
reinforcement equals the tensile strain in the soil in the
direction of the reinforcement.
Between the most critical plane and the wall face the
principal stress directions are approximately vertical and
horizontal, as indicated in Figs. 8 and 11. In this region
the reinforcement is aligned closely with the direction of
principal tensile strain in the soil.
6.2 Compatibility curve
The link betweeen the mobilised soil frictional resistance and
the mobilised reinforcement force can be shown on a
compatibility curve, Fig. 20 (Jewell, 1985).
In terms of required and available forces, equilibrium can
occur when the reinforcement mobilises sufficient available
force ~o satisfy the requirements for equilibrium, the
required force in the soil. This is where the two curves
intersect, Fig. 20c.
A smaller unit of reinforced soil could also be examined with
a compatibility curve in terms of the local required and
available stresses, rather than overall forces.
The compatibility curve relating overall required and
available forces in Fig. 20c is a direct illustration for
propped wall construction. The starting point of zero tensile
strain and unstressed reinforcement on the left axis is where
the fill and the reinforcement have been placed, but the prop
is pr0viding all the required force. The required force is
transmitted and distributed to the soil through the stiff
facing.
 393

On removal of the prop force there is a lack of equilibrium

and the soil must strain to allow the reinforcement forces to
be generated. As the soil strains it mobilises greater
frictional resistance, thus reducing the required forces. The
available reinforcement force increases as the reinforcement
strains with the sail, until equilibrium is established. The
overall required force is that which occurs on the most
critical surface through the toe.

:-1obilised A.xia 1
frictional force
resistance
_ :l.esponse to long
term loading

10 (7.) 10 (%)
Tensile strain Tensile strain
a. Soil characteristics b. Reinfo.rcement characteristics
(Isochronous curves)

Overall • Equilibrium point

re:quired
and
available -- dense sand
forces
_ - _ loose sand

10 (%)
Tensile strain
c. Compatibility curve

Fig. 20. Soil and reinforcement characteristics which may be

used " to indicate strain compatibility. (Jewell, 1985)

Stress paths
Another difference between propped and incremental wall
construction is illustrated in Fig. 21. As indicated above,
for a propped wall the soil will experience increasing
vertical stress during construction and the horizontal stress
is provided by the stiff, propped facing. Lateral strains
will be small if the propping is stiff, and the stress path RP
will lie close to the K, line, Fig. 21a. On removal of the
prop, the horizontal stress in the soil will reduce as the
reinforced soil moves into equilibrium straining laterally to
a higher st~ess ratio, PO Fig. 21a.
394

With incremental construction the reinforcement must provide

equilibrium at each stage. As the vertical stress increases
so does the horizontal stress required to maintain
equilibrium. Since the reinforcement is providing the
horizontal stresses, the soil will experience increasing
lateral strain throughout the construction process, therby
mobilising increasing frictional resistance. The stress path
in the soil for incremental construction is indicated by the
line RQ in Fig. 21a.
What is required for an assessment of compatibility is the
relationship between the mobilised frictional resistance and
the principal tensile strain for the soil following stress
paths similar to RPQ and RQ, under plane strain conditions,
Fig. 21a.

(b)

Fig. 21. Stress paths for reinforced soil for propped and
incremental wall construction.

6.3 Stress path in direct shear tests

The direct shear test is the most common plane strain test for
measuring the shearing resistance of granular fills. The
stress path in a direct shear test increases both the mean and
shear stresses in the soil up to the peak shearing resistance.
Because the principal axes in the soil rotate during a direct
shear test it is convenient to represent the test in terms of
the mean normal stress s and mean shear stress t. The stress
path UV (up to peak stress ratio) in a conventional direct
shear test is shown in Fig. 21b.
Although the stress path is steeper than that experienced by
the soil in an incremental wall, the direct shear test is as
close , as could be hoped for in a standard laboratory test.
The direct shear test is unsuited to modelling the unloading
stress path RPQ for a propped wall, Fig. 21. The most
suitable plane strain test would be in a biaxial apparatus,
with K, consolidation of the sample followed by reduction in
the minor principal stress.
 395

The unloading stress path is, of course, almost identical to

that in laboratory model tests on "active" retaining walls,
where the wall is relaxed away from a bed of soil. The
results of "active" wall experiments show that on an unloading
stress path the soil shearing resistance can be mobilised with
less lateral strain than on equivalent loading stress paths;
see the well known results in Lambe and Whitman (1968), for
example.
6.4 Strain softening
The stress paths in Fig. 21 are shown up to the peak stress
:-atio only. Continued shearing beyond the peak stress ratio
~n compact granular soils causes a gradual reduction in
shearing resistance, which levels off only when the critical
state has been reached at large strain.
In unreinforced dense granular soil the strain "post peak"
will concentrate along discrete shear surfaces if at all
possible. This is natural because once the shearing
resistance has begun to reduce locally in the deforming soil,
there is less resistance than in the adjacent soil so further
shearing will occur preferentially in the already weakened
soil. This mechanism of concentrated straining in a
relatively thin or discrete band through the soil allows the
overall shear strength to reduce with relatively small
deformations at the soil boundaries.
This is the traditional concern in the design of soil slopes
and walls where there may be non-uniform mobilised shearing
resistance through the soil. In this case the designer
anticipates that soil in a zone of relatively high strain
could be rapidly losing shearing resistance with small
deformations on discrete shear surfaces before the soil
elsewhere has strained sufficiently to mobilise the design
shear strength. The solution is to adopt relatively low
values of mobilised shearing resistance for the soil, or high
factors of. safety.
Local strain softening would also occur in reinforced soil in
the unreinforced soil behind the reinforced zone, or between
the soil and the reinforcement during outward sliding or when
the bond limit is reached between reinforcement and the
adjacent soil.
However, in the main body of a reinforced soil wall (where
bond is not limited) it seems unlikely that the mechanism of
local strain softening would occur. The reason is that as
long as there is available bond between the reinforcement and
the soil any local strain along a discrete shear surface
through .the reinforced soil would locally mobilise additional
reinforcement force which would re-establish equilibrium.
Between the most critical mechanism and the wall face there
are a family of parallel potential slip surfaces which are all
equally critically loaded by the soil self-weight, as was
shown in Fig. 4. If the deformation mechanism is considered
as slip along these parallel surfaces, slip along one
396

particular surface would locally mobilise additional rein-

forcement force and hence increased shearing resistance; the
adjacent parallel surfaces would then be more critical and
deformation would occur next along one of them in preference.
The process would then be repeated.
When designing for relatively extensible reinforcement,
therefore, it is the detail of the strain in the soil during
strain softening that is of interest, rather than the
deformation along a discrete shear surface through the soil.
It is the latter measurement that is recorded at the
boundaries of most laboratory tests "post-peak".
6.S Data from the simple shear test
The Cambridge University Simple Shear Apparatus is _a plane
strain test in which the central portion of the soil sample is
kept uniform thereby allowing detailed direct measurement of
the stress and the strain in the soil even during strain
softening, Stroud (1971). The stress path in a simple shear
test under a constant vertical stress is similar to the direct
shear test.
The finding from simple shear tests on dense sand is that the
stress ratio first increases rapidly with relatively small
strain, and then stays at a high level while considerable
strains develop in the soil. This is illustrated by Stroud's
simple shea,r test results on dense Leighton Buzzard sand in
Fig. 22. (Stroud, 1971). The stress ratio only gradually
reduces towards the critical state stress ratio.

0-8
tis
0-7

0-6

0-5

(tis )0 o-t. -L: v

a-at.
0-2
0-02
a-I

a
0-10 0-20 L:y

Fig. 22 Simple shear test results on dense Leighton Buzzard

sand (Sroud, 1971).
 397

The reason for this behaviour is that the critical state

shearing resistance is reached only at the critical state
specific volume in the sand (Schofield and Wroth, 1968). Very
considerable volume expansion is required in initially dense
sand to reach the critical state specific volume. This is
what causes the dilation when dense sand is sheared. Thus
large tensile strains develop in dense soil as it shears
towards the critical state.
This feature of behaviour for granular soils is well
illustrated by Stroud's simple shear data plotted in Fig. 23
(Stroud, 1971). The data are for three dense and three loose
samples, and the mobilised shearing resistance (tis) is shown
plotted against the specific volume in the soil. The specific
volume v 1 has been normalised to allow for different mean
stress levels in the tests. What is clear from the dat'a is
the very rapid increase in shearing resistance in the dense
sand which occurs with little volume change, followed by a
rather gentle reduction in shearing resistance accompanied by
considerable volume expansion (and tensile strain) as the soil
strains towards the critical state.

t 0,8
tIs
0·7

0·6 • Critical
state

0-5

0·4 Dense tests Loose tests

1·60 1·70 1·80 1·90

Fig. 23. The results of simple shear tests on sand plotted in

terms of stress ratio versus (normalised) specific volume.
(Stroud, 1971).

Similar detailed internal observation of the stress strain

behaviour of sand has been made in the Imperial College Hollow
Cylinder Apparatus (HCA) (Symes, 1983). The tests show the
same pattern of behaviour for dense sand on similar stress
paths; rapid mobilisation of shearing resistance at relatively
low strain, followed by continued straining with an only
gradual reduction in stress ratio.
398

Tests on dense sand in the HCA were also carried out on

unloading stress paths relevant to propped walls, and these
show 1 as anticipated 1 that less strain develops to mobilise
the shearing resistance of the soil on unloading stress paths
than on loading stress paths (at similar mean stress and
specific volume).
6.6 Mobilised shearing resistance versus soil tensile strain
For plane strain design most of the soils data comes from
unsophisticated direct shear tests. The triaxial test does
not represent the correct loading conditions. The plane
strain equivalent to the triaxial test l the biaxial test l
which could provide plane strain data is not widely used.
For design against a collapse limit state the difficulty of an
unsophisticated plane strain test is overcome by making a
conservative estimate for the shearing resistance of the soill
and allowing suitable safety margins.
A calculation of serviceability 1 however 1 requires
stress-strain properties for the soil. If there are no
stress-strain data on relevant stress paths 1 then the only
alternative for a prediction of compatibility is a simple
model that captures the main features of the soil behaviour.
The simple model for sand described below allows direct shear
test data to provide an approximate relationship between the
mobilised soil shearing resistance and the soil tensile
strain.
Simple model for sand
The soil surrounding an expanding cavity experiences large
strain. An analytical solution for cavity expans~on requires
a relationship between the radial displacement of the cavity
and the expansion pressure applied to the soil. Hughes et al
(1977) solved the problem of cylindrical cavity expansion in
frictional 1 dilating granular materials by modelling the
behaviQur of the sand as an elastic plastic material with the
following characteristics.
The elastic behaviour is given by the shear modulus G and the
drained value of Poisson's ratio v. When the sand reaches the
peak stress ratio (t/s)p it continues to deform with a constant
angle of dilation 'iJ and at a constant stress ratio (t/s)p.
For application to reinforced soil compatibility 1 the model
can be simplified one step further. Stroud (1971) observed
that the elastic volume change in dense sand was negligibly
small unless the mean stresses were changing substantially
along the stress path. If elastic volume change is ignored
the material response to shear from an isotropic stress state
can be summarised as shown in Fig. 24a l in terms of
(14)

and
 399

(15)

where dEl and dE3 are the incremental major and minor principal
strains.

tIs tIs

I
I
: sin +p
1/ 2(G/s)
£3

(£1+£3) £1

(a) (b)

Fig. 24 Simple elastic plastic model for dense sand (after

Hughes et al 1977).

The material behaviour is shown plotted as a function of the

tensile strain in the soil in Fig. 24b. The peak stress ratio
in the SOt I is reached with a mobilised tensile strain
() sinq'Jp
E3 poak = 2(G/s)
(16 )

For soil sheared from an initial stress ratio, the tensile

strain can be expressed in terms of the change in stress ratio
LI(lls) to the required mobilised strength, so that

( ) LJ(t/s)
E3 = 2(G/s)
(17 )

Shear mqdulus for sand

The elastic shear modulus for sand depends on the sand
density, the mean pressure and the strain amplitude. There
are no well accepted relationships available for the
400

evaluation of the shear modulus for sand in terms of other

standard material properties. (See Wroth and Houlsby, 1985,
for example).
An approximate estimate for the shear modulus could be made
from a direct shear test. One set of assumptions to determine
(Gis) would be:

(1) assume that there is uniform shear strain through the

full depth of the sample for the elastic deformation
so that the shear strain y is

y = o(x)/ H ( 18)

where 6(x) is the measured shear displacement and H is

the thickness of the sand sample .
(2) estimate the change in stress ratio between the start
of the test and the mobilised stress ratio to which a
secant shear modulus is to be calculated;
(a) if the stress path is assumed to start from an
isotropic stress state and end at the peak shearing
resistance then the change in the stress ratio would
be

Ll(t/s)=sin¢p (19)

so that the shear modulus is

sin¢ p
G/ s = -----'-- (20)
(o(x)/ H)

(b) an alternative, which allows for the rotation of

principal axes at the beginning of a direct shear
test, would be to use the shear displacement and the
change in stress ratio from the point of zero rate of
change of volume to the desired mobilised stress
ratio. The zero rate of volume change occurs at
about the critical state stress ratio which would
give

(sin¢p-sin¢cv)
G/s=~---"--------""-'- (21)
(o(x)'/H)

where 6(x)' is the relevant shear displacement, and ¢p

the mobilised angle of friction.

The use of this soil model and the equations presented above
is illustrated in the companion paper for the prediction of
strain compatibility in the RMC trial wall.
 401

REINFORCEMENT MATERIAL PROPERTIES

7.1 Relationship between force and extension

Isochronous load extension curves for the reinforcement
material are a suitable way to represent the influence of time
on the material properties of the reinforcement. The material
properties were presented in this way for the RMC trial, see
Bathurst and Jarrett (1986).
One set of isochronous curves applies at one temperature and
for one chemical environment and state of mechanical damage
for the reinforcement. For design purposes it is possible to
make conservative assumptions taking the highest likely
temperature and the worst likely combination of soil
environment and mechanical damage that might occur in the
reinforced soil wall.
Isochronous curves alone may be insufficient to allow accurate
estimates for stress relaxation, which might be important for
long lived structures. This area of material behaviour still
warrants more attention.
7.2 Stiffness at low strain
The working condition in most geotextile reinforced slopes and
walls to date is one with unexpectedly low reinforcement force
and deformation (see Yako and Christopher (1987) for recent
examples). There are additional factors which can affect
accurate predictions of working conditions with these low
reinforcement loads.
Firstly, there is often insufficient test data for the
reinforcement at low load.
Secondly, the pre-conditioning of the sample before testing
(ini tial load-unload cycles when setting up the sample) and
the gripping arrangements can influence the initial
load-extension properties.
Construction also inevitably introduces factors which are more
significant with relatively low reinforcement load, and hence
extension - slack in connections, misaligned reinforcement,
for example.
Another factor that is important at low loads is early changes
in the stiffness of the reinforcement with load level. The
"S" shaped load extension curve for some polyester materials
is a well known example - relatively low initial stiffness
which increases with increasing load (see Zanten (1986), for
example). Change in stiffness in the early part of tests can
become obscured by corrections to laboratory data to determine
the origin for strain.
402

BOUNDARY CONDITIONS

8.1 Introduction
The boundary conditions for a reinforced soil wall are likely
to affect the forces and displacements in the soil and the
reinforcement. In general the boundary conditions for walls
built on competent foundations are likely to reduce the forces
and displacements in the reinforced soil compared with the
assumed ideal boundary conditions.
In this paper it has been assumed that only the reinforcement
provides additional stabilising forces in the soil, and that
at the wall face the reinforcement tensile force is neatly
transferred into a uniformly varying compressive stress on the
soil surface. This may be thought of as ideal facing~ as the
face adds no kinematic restraint and is perfectly free from
interaction with the foundation.

8.2 Base boundary

The height of a wall is determined by the base boundary. The
base boundary defines the position of the most critical
surface through the toe, and divides the soil into the uniform
zone at the tace and the zone of continuously varying stresses
behind the most critical surface.
The rigidity and roughness of the base boundary affects the
stabili ty conditions in the zone behind the most critical
surface. The base boundary can be considered to have
roughness and strength provided by the foundation soil.
If the base boundary is not fully rough then potential failure
mechanisms with a horizontal portion along the surface of the
foundation may increase the required reinforcement forces
behind the most critical surface.
A base foundation with inadequate shearing strength to support
the ve~tical and shear loading from the wall at the foundation
level would inevitably directly affect the equilibrium in a
reinforced soil wall. This case of inadequate foundation
bearing capacity is not considered further.
If the base boundary is both rough and strong (which is most
common) it acts similarly to a plane of rough inextensible
reinforcement. r:r:he addi tional "reinforcing" reduces the
tensile strain in the soil near the base of the wall and hence
reduces the mobilised reinforcement force in the lowest layers
of reinforcement.
 403

8.3 Face boundary and connection with the base

The influence of the face boundary can be thought of in terms
of the face continuity and stiffness, and the connection with
the base.
The continuity of the face, the connection between panels, and
the stiffness of the face will provide kinematic restraint at
the face on the overall deformation in the reinforced soil.
An example was given earlier where the connection between
panels was anticipated to influence the magnitude of
construction induced movements by either allowing the panel to
translate or forcing it to rotate about the base, (section
5.2) .
A very stiff (unbending) continuous face, as in most propped
wall constructions, could act to locally increase. the
mobilised reinforcement force close to the face in the upper
part of a wall. This would occur if additional outward
deflection higher in the wall caused by the stiff face pulled
the reinforcement layers outwards. Local overstressing of the
reinforcement adjacent to the face caused by a stiff,
continuous face is probably not a problem in practice as
practical reinforcement spacing arrangements typically provide
more reinforcement layers than needed near the top of a wall.
The connection between the face and the base can also
introduce large stabilising forces into a reinforced soil wall
thereby reducing the reinforcement forces. Both vertical and
horizontal force can be transmitted from the foundation to the
wall.
The horizontal force would be transferred to the reinforced
soil as a normal stress at the face and would reduce the
required reinforcement force proportionally at that elevation.
The vertical force at the face would provide an upward shear
stress on the reinforced soil effectively making the front
boundary "rougher", in the sense of conventional active earth
pressure theory. This would locally reduce the vertical
stresses 'in the reinforced soil and hence the required
reinforcement force.
8.4 Summary on boundary conditions for practical walls
Most reinforced soil walls are built incrementally with
connected facing panels and on a competent foundation. These
boundary conditions reduce the reinforcement forces in the
lower reinforcement layers close to the base of the wall.
This comes from two main sources. Firstly, the rough, strong
base boundary acts like an additional substantial reinforce-
ment layer at the base of the wall. Secondly, the connection
between the face and the base can exert horizontal stabilising
stresses directly to the surface of the soil through the
facing. These stresses directly reduce the local required
reinforcement forces for equilibrium.
404

8.5 Sidewall friction

When small sections of wall are tested in the field or
laboratory the lateral boundary can also be important.
Usually a test section is confined between rigid parallel
boundaries and any movement in the soil results in shear
deformation between the soil and the stationary side boundary.
Any shearing resistance between the soil and the side boundary
mobilises boundary forces directly resisting the deformation
in the soil. So that in a Coulomb wedge calculation, for
example, a net side wall force acting in the opposite
direction to the implied soil movement should be included in
the equilibrium force balance.
Bransby and Smith (1975) used a detailed numerical method to
investigate the influence of side wall friction and wall
geometry (the ratio of wall height to width H/w) on the active
and passive earth pressure coefficients K. and Kp for
unsurcharged vertical walls. They found that the influence of
side wall friction on the active earth pressure coefficient
was quite small, a 14% reduction for a wall geometry H/w = 2
and a boundary friction 5.7".
For self-weight loading the influence of the boundary
roughness calculated by Bransby and Smith (1975) is
approximately uniform with depth. This is not the case for
surcharge loading where the side wall friction cumulatively
reduces the vertical stress in the soil. Side wall friction
can be more important for surcharge loading than for self
weight loading. Simple closed form analysis for side wall
friction is presented in the companion paper for self weight
and surcharge loading, where it is suggested that it has a
significant influence on the RMC trial wall.

CONCLUSIONS

A theoretical analysis for the behaviour of reinforced soil

walls has been described, and the derivation of key soil and
reinforcement material properties has been considered. The
ideas are illustrated by practical application in a companion
paper where they are used to make an analysis of the behaviour
of the RMC trial walls.

The main points and conclusions from this paper are summarised
below.
9.1 Equilibrium states
Two possible states of equilibrium have been presented for
reinforced soil walls, together with the analytical equations
describing them.
The ideal length equilibrium case is where the reinforcement
locally provides the required stress in the soil everywhere
that it is needed. The unreinforced soil behind the
reinforced zone does not load the reinforced zone.
 405

The truncated length equilibrium case is where equilibrium in

the soil is maintained with the reinforcement everywhere
carrying the maximum allowable force. In this case there are
unreinforced zones .. (behind the reinforced soil) that require
reinforcement stresses but do not contain reinforcement.
These come into an active state of equilibrium loading the
back of the reinforced zone, and the required reinforcement
force is thereby shed to the lower reinforcement layers.
It has been suggested that the two equilibrium cases are
likely to "bound" the actual state of equilibrium in a
reinforced soil wall.
9.2 Reinforcement force and deformation
The distribution of the reinforcement force in equilil;>rium
wi th the soil is described analytically for the two
equilibrium cases. The magnitude of the reinforcement force
depends on the reinforcement spacing. The concept of a
balanced reinforcement layout satisfying local and overall
equilibrium has been reviewed, and this determines the ideal
spacing arrangement.
The elongation in the reinforcement is an important component
of the outward movement in a reinforced soil wall. The
analytical calculation for the reinforcement elongations in
the two equilibrium cases is provided in non dimensional
charts which allow the deformation in the reinforced zone to
be directly calculated. The movement depends only on the
mobilised soil strength, the mobilised reinforcement force and
the reinforcement stiffness. The charts are applicable to any
reinforcement spacing arrangement.
The important finding (which is particularly emphasised in the
companion paper) is that the difference ~n the movement
calculated from the two "bounding" equilibrium cases is rather
small, and the analysis therefore predicts the deformation in
the reinforced zone to within close limits.
9.3 Incremental construction movement
Aligning facing panels during incremental construction above
the previously built facing which has already deformed causes
a moving datum for the position of the initially unstressed
reinforcement layers. This leads to cumulative additional
outward .movement that must be taken into account for each
layer. The additional movement can be thought of as the
deformation that would arise if the incremental layer were
built as an independent wall.
An analysis for incremental movement is presented. The
incremental movement should be added to the movement caused by
the reinforcement elongation to give the overall movement at
the face. Because the incremental movement is cumulative it
can significantly affect the outward movement in the upper
half of a wall.
406

Perhaps the most important distinction between an incremental

wall and a propped wall is that propping during filling
eliminates the moving datum and hence the incremental
construction movement. For the propped wall the outward face
movement is due only to the reinforcement elongation.
9.4 Strain compatibility
The tensile strain in the soil and the reinforcement provides
the link between the mobilised values of soil shearing
resistance and reinforcement axial force. This determines the
conditions for equilibrium in the reinforced soil. The
tensile strain in initially unstressed reinforcement cannot
exceed the tensile strain in the adjacent soil.
A compatibility curve can be used to estimate the equ~librium
state, and this is most conveniently carried out in terms of
the overall equilibrium of forces on the most critical
surface.
Biaxial tests, which would be the the most appropriate source
of plane strain data for the soil stress-strain characteris-
tics, are not widely used. In the more common plane strain
direct shear test the relationship between the mobilised
shearing resistance and the soil tensile strain is not
measured.
A simple elastic plastic soil model is proposed for use with
the compatibility curve which allows an estimate to be made of
the compatible equilibrium strain and mobilised shearing
resistance in a reinforced soil wall. A method is given for
estimating the relationship between the soil strength and
tensile strain from a direct shear test on sand. The
application of the method is illustrated in the companion
paper.
9 • 5 Summary
The theory and analysis in this paper provides a complete
method'for estimating the serviceability of a reinforced soil
wall. Two solutions can be derived which are expected to
"bound" the actual equilibrium. Non dimensional charts and
equations are presented which allow the reinforcement force
distribution along the layers (the maximum force distribution
in the layers is the same in both cases) and the face
deformation due to the reinforcement elongations to be
calculated. In practice, the difference between the predicted
movements for the two "bounding" equilibrium cases is
relatively small giving a rather precise prediction for the
likely movement of a reinforced soil wall.
Separate and additional movement due to incremental
construction of a reinforced soil wall has been identified.
This provides the major distinction between an incremental
wall where it occurs and a propped wall where it does not
occur. The calculation for the construction induced movement
 407

is very much less precise, although the magnitude of movement

is shown to be significant, particularly higher in the wall.
An analysis is given for incremental construction movement.

Acknowledgements
Aspects of the reported study were completed while the author
was supported by the Royal Society/SERC industrial fellowship
scheme. The author is grateful to the Soil Mechanics Group at
the Department of Engineering Science, University of Oxford.,
for stimulating support, to Guy Houlsby for the many
discussions on the behaviour of sand, and to George Milligan
for thoroughly reviewing the manuscript and making many useful
suggestions.

References
Bathurst, R.J. & Jarrett, P.M. (1986). Class A predi ction
exercise for reinforced earth walls. Bulletin No.1 for NATO
Advanced Research Workshop, Application of Polymeric Rein-
forcement in Soil Retaining Structures, Royal Military
College, Kingston.
Bransby, P.L. & Smith, I.A.A. (1975). Side friction in model
retaining wall experiments. Journal of Geotechnical Engineer-
ing, ASCE GT7, July, 615-632.
Hughes, J.M.O., Wroth, C.P. & Windle, D. (1977). Pressureme-
ter tests in sand. Geotechnique 27, 455-477.
Jewell, R.A. (1985). Limit equilibrium analYSis of rein-
forced soil walls. Proc. 11 Int. Conf. Soil Mech. Fdn Engng,
San Francisco, Vol 3, 1705-1708.
Jewell, R.A. (1987). Analysis and predicted behaviour for
the Royal Military College trial wall . Proc. NATO Advanced
Research ~orkshop, Application of Polymeric Reinforcement in
Soil Retaining Structures, Martinus Nijhoff.
Jewell, R.A., Paine N.P. & Woods R.I. (1984). Design methods
for steep reinforced slopes. Proc. Int. Conf. Polymer Grid
Reinforcement, London, 70-81.
Jewell, R.A. and Wroth, C.P. (1987). Direct shear tests on
reinforced sand. Geotechnique 37, No. I, 53-68.
Lambe, T.W. & Whitman, R.V. (1968). Soil Mechanics. John
Wiley, New York .
Stroud, M.A. (1971). The behaviour of sand at low stress
levels in the simple shear apparatus. PhD Thesis, University
of Cambridge.
Symes, M.J.P.R. (1983) . Rotation of principal stresses in
sand. PhD Thesis, Imperial College, London.
Schofield, A.N. & Wroth, C.P. (1968). Critical State Soil
Mechanics, McGraw Hill, London.
408

Wroth, C.P. & Houlsby, G.T. (1985). Soil mechanics -

property characterisation and analysis procedures. Proc. 11
Int. Conf. Soil Mech. Fdn Engng, San Francisco, Vol 1, 1-55.
Yako, M.A. & Christopher B.R. (1987). Polymerically rein-
forced retaining walls and slopes in North America. Proc.
NATO Advanced Research Workshop, Application of Polymeric
Reinforcement in Soil Retaining Structures, Martinus Nijhoff.
Zanten, R.V. Van (ed.) (1986). Geotextiles and Geomembranes
in Civil Engineering, Balkema, Rotterdam.
REINFORCEMENT EXTENSIBILITY IN REINFORCED SOIL WALL DESIGN

Rudolph Bonaparte, GeoServices Inc. Consulting Engineers, USA

Gary R. Schmertmann, Graduate Student, University of California,
Berkeley, USA

1. INTRODUCTION
There are many types of reinforcing materials and systems
available for the construction of reinforced soil walls. Of the many
types, the Reinforced Earth system developed by Vidal (1966) in France
has predominated and has provided the basis for most theoretical and
empirical knowledge of the behavior of reinforced soil walls. The
Reinforced Earth system has a number of distinguishing characteristics
that include:
• steel reinforcing elements that have tensile modul i on the
order of 2 x 10· kPa (3 x 10 7 lbs/in2);
• reinforcing elements that are discrete strips, approximately 50
mm (2 in.) wide and 5 mm (0.2 in.) thick; and
• concrete facing (skin) elements that ~an individually undergo
limited translation and rotation in response to movements in
the reinforced fill or settlements of the foundation soils.
More recently, reinforced soil walls have been constructed with
geosynthetic, reinforcement and various facing elements. The two types
of geosynthetics commonly used in reinforced soil wall construction
are geogrids and geotextiles. Geosynthetic soil reinforcement systems
have distinguishing characteristics that include:
• polymer rei nforci ng el ements that have tensil e modul i on the
order of 1 x 10 6 kPa (1.5 x 10 4 lbs/in2);
• reinforcing elements that are continuous or semi-continuous
sheets with thicknesses in the range of 1 to 5 mm (0.04 to
0.20 in.); and
• a variety of possible facing (skin) elements including concrete
panels, timbers, and geosynthetics.
It is clear that the Reinforced Earth system and geosynthetic
reinforcement systems incorporate fundamentally different types of
reinforcing elements: steel strips versus polymer sheets. While the
nature of the reinforcing elements are different, the most commonly

409

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 409-457.
© 1988 by Kluwer Academic Publishers.
410

used procedures for analysis and design are similar. These procedures
derive largely from research on Rei nforced Earth and other steel
reinforcement systems. There prevails an inherent assumption that
analysis and design procedures developed for steel reinforcement are
also applicable to geosynthetic reinforcement. This assumption is
questionable.
This paper will examine differences in the behavior of soil walls
reinforced with steel strips or other relatively "inextensible"
materials (such as steel grids) and soil walls reinforced with
relatively "extensible" geosynthetic materials (geotextiles or
geog ri ds) • Thi s compari son will be made for the idea 1 case of a
frictionless wall facing, braced during construction, and free to
rotate about its toe after construction. Wall systems and
construction procedures causing deviations from this "ideal" case will
be discussed. Analysis and design procedures for reinforced soil
walls will be reviewed and conclusions will be drawn on the procedures
most appropriate for use with geosynthetic reinforcement. These
procedures will be used to conduct parametric studies and to predict
the behavior of the instrumented geogrid-reinforced soil walls
constructed at the Royal Military College (RMC) in Canada.
2. COMPARISON OF INEXTENSIBLE AND EXTENSIBLE REINFORCEMENT
There are two basic differences between steel and geosynthetic
reinforcing elements:
• as stated in the introduction, the tensile moduli of steel
reinforcement and geosynthetic reinforcement are vastly
different: the ratio, Es/Eg, is on the order of 100 to 1000
(Es = modulus of elasticity of steel and Eg = low-strain
tensile modulus of polymers used in geosynthetics); and
• as shown by design experience, the volume of steel in a steel-
reinforced soil wall is much smaller than the volume of polymer
in a comparable geosynthetic-reinforced wall (0.02% to 0.05%
for steel reinforcement compared to 0.2% to 0.5% for
geosynthetic reinforcement).
The consequences of these differences are discussed below.
2.1 Definitions
McGown et ale (1978) originally defined inextensible and
extensible reinforcements (inclusions) as follows:
• inextensible reinforcements (inclusions) are those that "have
rupture strains which are less than the maximum tensile strains
in the soil without inclusions, under the same operational
conditions"; and
• "extensible reinforcements (inclusions) are those that have
rupture strains larger than the maximum tensile strains in the
soil without inclusions, under the same operational
conditions".
 411

These definitions are difficult to interpret since the word

extensible relates to the stress-strain response of a material rather
than to the material's rupture strain. A better understanding of the
meaning of the two terms, extensible and inextensible, can be
obtained by comparing the horizontal strain in an element of
reinforced soil subjected to a given load, to the strain required to
develop an active plastic state in an element of the same soil without
reinforcement:
• inextensible reinforcement is reinforcement used in such a way
that the tensile strain in the reinforcement is significantly
less than the horizontal extension required to develop an
active plastic state in the soil; and
• extensible reinforcement is reinforcement used in such a way
that the tensile strain in the reinforcement is equal to or
larger than the horizontal extension required to develop an
active plastic state in the soil.
As will be shown subsequently, steel rei nforcement meets the
criterion for inextensible reinforcement in most practical
applications. It will also be shown that currently available
geosynthetic reinforcing materials meet the criterion for extensible
reinforcement in almost all practical applications.
Two extreme cases of extensible and inextensible reinforcement can
be considered: (i) an "absolutely" inextensible reinforcement which
is so stiff that equilibrium is achieved at virtually zero horizontal
extension (Ko conditions theoretically prevail in the reinforced soil
mass if compaction induced stresses, soil arching, and other secondary
effects are ignored); and (ii) an "absolutely" extensible
reinforcement which has such a low modulus that virtually no tensile
forces are introduced to the soil mass at the strain required to
develop an active plastic state (Ka conditions theoretically prevail).
2.2 Stresses and Strains in Unre1nforced Soil Element
A sma"l element of compacted granular soil is considered at some
intermediate point in the construction of a retaining wall with a
frictionless, rigid facing and no reinforcement (Figure la).
Initially, the element of soil is in an at-rest state of stress (point
A in Figures 2a and 2b). As successive fill levels are placed, the
major and minor principal stresses acting on the soil element
increase. Several possible stress increments associated with wall
construction are shown in Figure Ib and corresponding stress paths are
shown in Figure 2a. Stress increments include: (i) compression
loading; (ii) compression unloading; (iii) proportional loading; and
(iv) general incremental loading. It is instructive to relate these
stress increments to the soil element "A" in Figure la. If the wall
facing is braced during fill placement, the minor and major principal
stress increments (ignoring compaction stresses) will be proportional
and their ratio will be equal to Ko (line segment AB in Figure 2a).
If the braces are removed after construction and the facing is free to
rotate about its base, 0 3 will decrease by ol·(K o - Ka) along a
compression unloading stress path (line segment Be in Figure 2a) as
the fill yields into an active plastic state.
412

(8)
FACING
r - - - -· --- --
~_ _ __ _ _____ ~SUBSEQUENT FILL LEVELS
I /
1------- ---
H
~------- _ EXISTING FILL LEVEL
o
~ SOIL ELEMENT
AT INITIAL
CONDITION A
;;)7 ............. "71" o:::::;c;:::<: OJ

'
(b)

Qa
INITIAL COMPRESSION COMPRESSION
CONDITION (A) LOADING (AC) UNLOADING (SC)
I~1a1

03
O a3
6a3
03 a3
~3

a, a,
t l!. a,
a3 /{a, +60,) = K.
PROPORTIONAL GENERAL INCREMENTAL
LOADING (AB) LOADING

O
J~,a, O~~,O,
~ ..0'3 ~3 "£03
~ ~ ~ ~
a 1 a,
t60, t6a1
(a3+603)/(C11+6a1)=Ko l!.a3 /l!.C1, 4: CONSTANT

Figure 1. Possible stress states for soil element A in a reinforced

soil wall: (a) soil element "A"; (b) stress states.
 413

(a)

Ka

GENERAL INCREMENTAL
LOADING STRESS PATH

(b)
C C C
I l ~

B GENERAL INCREMENTAL LOADING

STRESS-STRAIN CURVE

A •

o 2

Figure 2. Idealized stress-paths and stress-strain curves for stress

states in Figure 1: (a) stress paths; (b) stress-strain
curves.
 414

( a). 2.'

I I
v--
2.0
T
II v
L...- v- r-t- !:c:: -I-<
i V -r---- t-- to-
I., 11 v
/
"
~~ I ''''-
"'. . , ""
Preshtar hilult

.... ,,.... ,-,. ,. ..,."..•. ,....". ...

TestT)'Pe
1 b
'"
.... ''''....

.., ··
b'" If ~ ....... ,.U ,.~

, ~ ,. ~ 11.' ,~ CO"

.... ,.
c-pr........ • .19 '-51 U%"
..
...
1.0
0 c.."'nsioII ........ 0.51 '.51 U'
,."
.~ ~ ,. ~ I~

I. 6
eo.,ressioll ......
~1IIIIDIIIinf:
I.~
1.11
I."
I.~
I.~
1.U .~
,.»
,.~ ...
411f"

" '" -- .,.''""

X( httlrsiollllMdiltt O,~l I.M ,~ 0.55
,."
· ....... ,-.... ,. ,.
bltMiololDlllille: 0.51 0.55

"" ,."
.~ O~5 ,~

.04

.'
, .~ .~ It.S1

10."
bt.lISiDn ...... I... 0.55 I.~ , .~ CI'
)( ~ I .• I ." •. U

00
10 12

(b),
(c).

I LKP
-- --
I
Test 3
f,,-r,:--

1\ • I

L1 4

~t:-= ~
../
3

-40L-...L-..l.2----!-3-~4 -2 2
~("'cm')
1

Ka._
_K,
0

*-
-10 -$ 0 +5 +10
HoriIontM (~), f.h

Fi.gure 3. Stress-paths and stress-strain curves for dense sand

specimens in a triaxial tests: (a) stress-strain curves;
(b) stress and strain paths; (c) plot of lateral earth
pressure coefficient, K, versus horizontal strain derived
from test results for compression unloading and extension
loading. (From Lambe and Whitman, 1968.)
 415

In contrast to the above case, most reinforced soil walls are

built with facing elements that can deform incrementally as the wall
is built. Consider soil element "A" for this case. Initially, the
soil element will follow the Ko stress path (again ignoring compaction
stresses). As lateral stresses induce deformation of the wall face,
however, ~o. < Ko~o,. Eventually, 0 < ~o. < Ka~o,. The stress path
for this general incremental loading must fall between the stress
paths in Figure 2a for compression loading (line segment AC in
Figure 2a) and Ko loading followed by compression unloading.
Stress-strain curves associated with the stress paths in Figure 2a
are shown in Figure 2b. The curves in Figure 2b have been idealized
from data such as that shown in Figure 3. Starting from Ko
conditions, the major principal strain for soil element "A"
corresponding to the stress path AB + BC is 0.5% or less (Lambe and
Whitman,1968). Assuming E." - £, (i.e, E2 = 0), Eh = £. < 0.5% for
an active plastic stress state. For compression loading, Eh .. 2% to
3%. ' By deduction, for general incremental loading Eh .. 1% to 2%.
2.3 Stresses and Strains in Reinforced Soil Element
It was just shown that for an element of unreinforced compacted
granular soil, Eh < 0.5% for a compression unloading stress path, and
Eh .. 1% to 2% for a general incremental loading stress path. In this
section, the mobilized reinforcement forces at these strain levels are
investigated.
An element of compacted granular soil is considered to be
initially in an at-rest (Ko) state of stress. Plane-strain
deformations are assumed and compaction induced stresses are
neglected. Principal stress rotation due to the horizontal tensile
stresses is also neglected and thus the ratio of minor to major
principal stress is taken as:

where 0v is the vertical effective stress and 0h = horizontal

effective stress.
To evaluate the influence of reinforcement extensibility, the
relationship between K and horizontal strain, Eh, will be defined in
the subsequent analyses using a hyperbola. (Hyperbolas have been
widely used in soil mechanics to describe the shape of stress-strain
curves (e.g., Kondner, 1963; Duncan and Chang, 1970).) The hyperbola
used to define the K-Eh curve in Figure 4a is given by:
Ko - Ka
Ko - K = - - - - - (2)

where Ko is assumed to be given by Jaky's (1944) equation:

416

K 0.5
FILL PROPERTIES (a) .
K
</>'- 35°
Q ~m= 20kN/m 3
0.4

0.3
K
•
0.2 0.1 0.2 0.3 0.4 0.5
0 £.h (%)

1 Z(m)

10
2 ( b).

,.. "20
4
C\I
E
.....
Z
.lI: 30

I-
e.
GIS
s:. 40
b
<!h ~KIS'Z
8

eo

10

eo

10

80

Figure 4. Graph for estimating equilibrium strains in an ideal

element of reinforced soil subject to a compression
unloading stress path: (a) hyperbolic curve of
coefficient of lateral earth pressure, K, versus
horizontal strain Eh; and, (b) relationship between
horizontal effective stress, 0h, and Eh. (See Figure 4
for definitions of Ss and Sg.)
 417

Ko = 1 - sin $ (3)

and Ka corresponds to the active Rankine state of stress, given by:

1 - sin $
Ka = - - - - (4)
1 + sin $

In Equation 2, K = Ka at infinite strain. The parameter "h"

controls the relationship between (Ko - K) and €h. The value of
"h" can be estimated from test results or backcalculated from a set
of assumed conditions. To backcalculate "h" it is assumed that
(Ko - K) = 0.95 (Ko - Ka) at 0.5% strain for the compression
unloading stress path. In other words, at 0.5% strain, 0h is a~sumed
to have decreased from Koov to a value just slightly larger than
Kaov. The hyperbolic curve shown in Figure 4a is related to a
granular soil with $ = 35° and y = 20 kN/m" (127 lb/ft"). For this
soil, Ko = 0.426, Ka = 0.271, (Ko - Ka) = 0.155 and h = 2.6 x 10- 4 •
Figure 4b shows the relationship between 0h and €h obtained using
K values, calculated as indicated above, in the equation oh = Kyz,
with Y = 20 kN/m 3 • If it is assumed that the reinforcement tension
just balances 0h, the curves in Figure 4b represent the reinforcement
force per unit soil area required to maintain equilibrium in the soil
element. With this interpretation, Figure 4b can be used to estimate
the strains in a soil element with reinforcements having various
tensile stiffnesses.
To lend practical value to these estimates the steel and
geosynthetic reinforcement configurations shown in Figure 5 are
considered. Figure 5a represents a standard Reinforced Earth panel
with four steel reinforcing elements (strips). Typical dimensions of
the steel strips are given in the figure. From the information
provided in Figure 5 a reinforcement stiffness per unit soil area
(with the soil area taken normal to the direction of the reinforcing
element) can be defined for steel strips as follows:

Ss = Es • N • a • b/A (5)

where: Ss = steel reinforcement stiffness per unit soil area (kPa);

Es = tensile modulus of the reinforcement (kPa); N = number of
reinforcing elements (strips or layers) per wall facing panel; a and b
= dimensions of the reinforcing elements (strips or layers) as shown
in Figure 5. For a Reinforced Earth structure with E = 2 x 10· kPa
(29 x 10· lbs/in 2), N = 4, a = 4 mm (0.16 in.), b = 50 mm (2 in.), and
A = 2.25 m2 (24.5 ft 2), Ss = 70,000 kPa (1.5 x 10· lbs/ft2).
For geosynthetic reinforcement, the tensile modulus, Eg , and
thickness, a, are usually combined into a reinforcement stiffness per
unit width, J g (kN/m). In this case, Equation 5 becomes:
 418

Typical Values
(a>. a 4 to 6 11m
b 50 to 60 11m
N 4

f
Es 3 X 10 7 kPa
E

-
"! Stiffness Pe r Un ita rea
Ss = Es N a b/A
I (A = panel area)
I. 1.3m

1.6 m

Typical Values

(b) • a 1 to 4 11m
b 1m
( N 2 to 4
500 kN/m
b
.1 Jg
Stiffness Per Unit Area
Sg = J g N b/A

Typical Value
( c).
sv = 0.5 m

I:: Stiffness Per Unit Area

Figure 5. Equation for estimating the reinforcement stiffness per

unit soil area, S (kPa): (a) steel strip reinforcement
(Ss) with a segmental facing panel; (b) geosynthetic
reinforcement with a segmental facing panel (Sg); and (c)
continuous layers of geosynthetic reinforcement. The unit
soil area is normal to the direction of the reinforcing
elements.
 419

5g = J g • N • b/A (6)

Equation 6 is valid in the case of geosynthetic strips

(Figure 5b). In the case 0·1' continuous layers, the reinforcement
stiffness is given by:
5g = Jg/s v (7)

where Sv is the vertical spacing between reinforcement layers.

To estimate 5 g for a typical geosynthetic reinforcement
application, the Tucson, Arizona retaining wall project presented by
Berg et ale (1986) is considered. In this project, continuous layers
of geogrid reinforcement were used with vertical spacings between
layers ranging from 0.30 m to 0.75 m (12 and 30 in.). The low-strain
secant J g of the geogrid used on the Tucson project is on the order of
600 kN/m (40,000 lbs/ft). For the case of Sv = 0.3 m (12 in.), 5g •
2,000 kPa (40,000 lbs/ft2).
Once 5 s or 5g is obtained, the required reinforcement force per
unit soil area, T (kN/m2), can be related to the reinforcement strain
as follows:
T =5 • Et (8)

where: 5 = 5s (for steel) or 5g (for geosynthetic) as defined by

Equations 5, 6, or 7; and Et = tensile strain in the reinforcement.
In Figure 4b, the line for 5s = 70,000kPa is related to steel
reinforcement and the line for 5g 2,000 kPa is related to
geosynthetic reinforcement. Equation 8 is represented by straight
lines in Figure 4b.
In this paper it is assumed that Et = Eh. The use of Eh = Et
implies that: (i) the presence of reinforcement does not locally
affect soil strains (which is not true for steel and may not be true
for geosynthetics); and (ii) there is no slippage between the soil and
reinforcement. (This is a good assumption for geogrids reinforcing
9ranu 1ar fill since the fill penetrates the ape rtu res of the 9 ri d.
The assumption may be less appropriate for geotextiles where there is
a greater likelihood of localized slip between the fill and
geotextile). The intersections of the T-Et lines with the 0h-Eh lines
in Figure 4b represent equilibrium states for the reinforced soil
element. These two sets of lines are analogous to the strain
compatibility curves presented by Jewell (1985b).
The ·strains obtained from Figure 4b have been replotted in
Figure 6. This figure includes results for a range of steel
reinforcement stiffnesses (5 s 70,000 to 300,000 kPa) and
geosynthetic reinforcement stiffnesses (5 g = 500 to 3,300 kPa). It
can be seen that for steel reinforcement, the equilibrium strains are
on the order of 0.01% to 0.1%. The range of strains in Figure 6 for
geosynthetic reinforcement is large due to the wide range of tensile
stiffnesses associated with currently available products. Even for
420

EQUILIBRIUM STRAIN, ~h & et (%)

0~~____0_·r5_______1~.~0_______1~._5_______2~._0______-12.5

000 kPa

e
RANGE FOR STEEL
I
I
I
8.= 70,000 kPa
Z(m) 10 I

Figure 6. Range of equilibrium strains in an element of reinforced

soil incorporating either steel reinforcement (5 s ) or
geosynthetic (59) reinforcement and following a
compression unloadlng stress path. (5 s and 5g are defined
in Figure 4.)
 421

the stiffest geosynthetics currently used, however, strains are on the

order of 20 times greater than those for steel reinforcement. For the
case of Sg = 2000 kPa (40,000 lbs/ft 2 ) , strains range from about 1% up
to about 2.5%.
2.4 Lateral Stresses and Mobilized Tensions in Reinforced Soil
The equilibrium strains obtained from Figure 4b have been used
with the hyperbolic K-€ curve of Figure 4a to produce Figure 7. The
derivation of the curves in this figure is as follows:
• Lateral stresses in the soil are given by:

(9)

• Lateral stresses are transferred to reinforcement:

(10)

• Since it is assumed that €t = €h:

( 11)

• Combining Equations 2, 9, 10, and 11 gives the following

equation which was used to generate the K - z curves presented
in Figure 7:
AS Ko - K
z =-- ( ) (12)
. Y K (K - Ka)
where: z = depth; A = parameter used in Equation 2 and linked
to the strain required for the soil to reach the Rankine active
state of stress; S = reinforcement stiffness per unit area; y =
unit weight of soil; Ko = at-rest lateral earth pressure
coefficient; Ka = active lateral earth pressure coefficient;
and K = lateral earth pressure coefficient at strain €h.
Figure 7 presents a plot of the mobilized reinforcement force per
unit area (normalized in terms of the lateral earth pressure
coefficient, K) required to keep a soil element at depth z in
equilibrium. Results are presented for both geosynthetic and steel
reinforcement for a soil with A = 2.6 X 10- 4 • Figure 7 is revealing
because it illustrates the role of reinforcement tensile stiffness on
the horizontal stresses that must be resisted by the reinforcement.
With geosynthetic reinforcement, the calculated lateral stresses are
virtually those corresponding to the active Rankine state, even for
relatively stiff geosynthetic reinforcement. Clearly, geosynthetics
fulfill the criterion given previously for extensible reinforcement.
 422

LATERAL EARTH PRESSURE COEFFICIENT. K

0.1 0.2 0.5

O,--------+--------+-----~~~~~--~t_----~

RANGE FOR K VERSUS DEPTH

GEOSYNTHETICS ASSUMED BY THE
COHERENT GRAVITY
PROCEDURE

2
Sg = 2,000 kP.
j---- Ko LINE REPRESENTS
'ABSOLUTELY'
I IN EXTENSIBLE
I REINFORCEMENT
I
FILL PROPERTIES
I
9\'=311° I
---I-~ RANGE FOR STEEL
lSm=20kN/m3

8- K LINE REPRESENTS 0

• 'ABSOLUTEL Y'
~
EXTENSIBLE
REINFORCEMENT I
f------/-- S. 70 , 000 kPa

10'

Z(m) 12

Figure 7. Calculated lateral earth pressure coefficient, K, for an

element of reinforced soil incorporating either steel (Ss)
or geosynthetic (Sg) reinforcement and following a
compression unloading stress path, Calculated values are
based on equil ibrium strains obtained from Figure 4. (Ss
and Sg are defined in Figure 5.)
 423

According to Figure 7, mobilized reinforcement forces and lateral

earth pressures are larger for steel than for geosynthetics. Figure 7
indicates that for steel, lateral earth pressures near the top of a
wall should be close to those associated with at-rest conditions. In
this case, steel reinforcement fulfills the criterion for inextensible
reinforcement. At depth, the lateral earth pressures tend toward
those assoc i a ted with act i ve condit ions. Based on the hyperbo 1 i c
curve in Figure 4a, steel reinforcement would need to undergo a
tensile strain on the order of its initial yield strain (Eg - 0.2%) in
order for the soil to reach an active plastic state. Since the steel
reinforcement used in Reinforced Earth walls is designed so as not to
exceed a tensile stress of about 0.55fy , where fy is the initial yield
stress of the steel, the reinforced fill shoula theoretically never
reach an active plastic state.
This last observation is interesting in light of the assumed
distribution of lateral earth pressure versus depth used in the
coherent gravity design procedure recommended by LCPC - SETRA (1979)
for the design of Reinforced Earth walls (also see Ingold, 1984 or
Jones, 1985 for a description of the procedure). In the coherent
gravity procedure, the lateral earth pressure coefficient, K, is
assumed to decrease from a value of Ko at the ground surface to Ka at
(and below) a depth of 6 m (20 ft). Based on the above discussion, it
is concluded that Ka stresses can only develop in a steel reinforced
soil structure if there is slip between the soil and reinforcement.
If there is no slip, a larger value of K should be considered for the
coherent gravity procedure for depths greater than 6 m (20 ft).
Figure 8 is similar to Figure 7 except that it shows the influence
of the K-Eh relationship on the calculated K - z curves. Results are
shown for Et = Eh = 0.5%, 1%, and 2%. As previously noted Eh = 0.5%
is assumed to be correspond to a compression unloading stress path and
Eh = 1% to 2% is assumed to correspond to a general incremental
loading stress path. It is clear that while the effect of stress path
(and thus Eh) is significant, it does not alter the conclusions of the
previous paragraphs.
Mitchell (1987) has presented results from incremental, plane-
strain finite element analyses conducted by Collin (1986) consistent
with the results presented above. The finite element model used by
Collin accounts for nonlinear soil stress-strain properties, including
dilatancy and hysteresis, independent facing units, soil-reinforcement
interface behavior, incremental construction processes, and compaction
induced stresses. The model was used to predict the behavior of 4.3 m
(14 ft) and 6.1 m (20 ft) high reinforced soil walls. The walls were
intended to model a steel bar-mat reinforced soil wall constructed in
Hayward, California. The geometries and properties of the walls are
shown in Figure 9. Numerical simulations were conducted for three
types of reinforcement: steel bar-mat (Ss = 112,000 kPa); welded wire
mesh (S5 = 310,000 kPa) and geogrid (Sg = 1400 kPa). Resul ts are
shown in Figure 9 along with measurements obtained from the Hayward
Wall. Inspection of this figure shows that for the steel reinforcement
both the numerical simulations and measured data indicate 0h
magnitudes in the range of Koov or greater, with 0v calculated using
the "Meyerhof (1953)" type stress distribution to account for the
lateral thrust of the retained backfill. Similar results were
424

LATERAL EARTH PRESSURE COEFFICIENT, K

0.1 0.2 0.5

GEOSYNTHETIC REINFORCEMENT /"

""",~y

Sg= 2,000 kPa

eh=t!?,
J-H-+-=."*",, STEEL REINFORCEMENT
2 0.5% S. = 70,000 kPa
1.0%
2.0%----!H
I--- Ko LINE
I
4 I
I

e. h = ~\
8
2.0%
1.0%
I------r-- 0.5%

FILL PROPERTIES
10
",'= 35 0
"lf m =20kN/m 3
z(m)

12

Figure 8. Influence of stress path on the calculated lateral earth

pressure coefficient, K, for an element of reinforced soil
incorporating either steel (Ss) or geosynthetic (Sg)
reinforcement: €h '" 0.5% corresponds to compression
unloading (Figure 2a) and €h = 1.0% to 2.0% corresponds to
general incremental loading.
 425

obtained from the numerical simulations of the welded wire mesh wall.
The results from the numerical simulations incorporating geogrid
reinforcement indicate 0h = Kao v , where 0v is the overburden pressure
(ov = yz). This result is consistent with Figures 7 and 8.
2.5 Influence of Construction
The preceding analyses were for single elements of reinforced
soil. These analyses can be extended to the behavior of a
geosynthetic-reinforced wall if it is assumed that the reinforcement
does not affect the stress state in the soil (i.e., the soil stresses
and reinforcement tensions are uncoupled).
For the case of an unreinforced wall that is braced during
construction, but free to rotate about its toe after the braces are
removed, the wall fill undergoes Ko loading followed by compress ·ion
unloading (stress path AB + BC in Figure 2a). If the wall face were
frictionless, an active Rankine state would theoretically develop.
This latter case corresponds to that of a flexible, frictionless
cantilever wall rotating about a fixed toe (Figure lOa). If the wall
contains inextensible reinforcement that makes a good bond with the
soil, the large tensile stiffness of the reinforcement will suppress
horizontal soil strains. In the limiting case of an "absolutely"
inextensible reinforcement, the horizontal direction becomes a zero
extension direction and the failure surface is as indicated in Figure
lOb (Bassett and Last, 1978). For the case of extensible
reinforcement, horizontal soil strains should be only partially
suppressed, as indicated by Figures 7 and 8. It is hypothesized that
the failure surface for this case will be close to that shown in
Figure lOa.
For the case of incremental wall construction the situation is
more complex. At depth, the soil will follow some general incremental
loading stress path (Figure 2a), eventually reaching an active plastic
condition. Near the top of the wall, the soil should be in a
subfailure state, with the soil stress path being closer to one of
proportional loading (AB in Figure 2a) than unloading (BC in figure
2a). The failure mechanism for this case is more complex than given
by Figure lOa. Suggestions have been put forth that the fail ure
mechani sm iss imil ar to one for a reta in i ng wa 11 deformi ng about a
hinged crest. There is some experimental support for this suggestion
for inextensible reinforcement systems (Juran and Schlosser, 1978) but
not for extensible reinforcement systems. Additional research is
needed in this area.
2.6 Conclusions for Geosynthetic Reinforcement
It is concluded that currently available geosynthetic materials
are extensible forms of reinforcement. This conclusion is based on an
analysis of the amount of horizontal extension required to induce an
active plastic state in an element of unreinforced soil. The amount
of extension is s.tress path dependent and, for a compacted granular
fill, ranges from less than 0.5% for compression unloading up to 1% to
2% for general incremental loading. The amount of mobilized
geosynthetic reinforcement tension at these low extensions is small.
Theoretically, local yielding of the soil element will occur before
 426

( b).
LATERAL EARTH PRESSURE (p")
(aL 500 1000 1500

6 Predicted by FEU
o Meosurt:d
~ g~fJ';~d \~r:e f~~ FEM
A. _ I(Q.Ov"bu'd .... P~e"ure
"b' A D a_ Ko.Bea.inq Preuu.e
GRAVELLY SAND C - I<O·(}v.,rburdcn Prenu'.
to.
o ", 60 - KO·0.,0";"9 P.usun

20'
¢>'~ 40.6°
g,IO
" ~

~\, ,'\ ' ~

~m =122 Ibo/f!3 I-
OR 0>
o ~ ,,'"
14' e:
,'
i015 0\ \ .. '"
e;
o , ,
o \ \ " It.

'.A,S C I\,
FOUNDATION 2 I 20
Su=1100IbO/~! l,lS'OR14'.
lSm=107 Iba/f!

( c).
14 Ft High Wall 20 Ft High Wall
DEFLECTION (in) DEFLECTION (in)
:3 2 1 '0
Jr--,r---'l2:'-'---r1--r-'/O 0 -::; 0
a VSL Bar tlot
o VSL Bar tAat " Welded Wire
• Weldep Wife o Geogrid 2-::;
o geognd o c 2..J o o •
o c. ..J
..: 0 4 -1
o O· 4 3= 0 -1
o lJ.. 0
6 ~
o 6 0
0...
o 0
8 \:i
o 8 I-
0 10~
o 2 0 f-

o 1~
lJ..
0 122
0 o • 0 0
o 0 • 12~ 14B:
0... o •
00' W 0 o • 16~
"'" 14° w
0...
0 o. 180
o o.
20

Figure 9. Predicted and measured performance of Heyward reinforced

soil walls: (a) wall cross-section and steel bar-mat
reinforcement layout; (b) measured and predicted lateral
earth pressures for different types of reinforcement; (c)
predicted wall deformations . Wall facing consiste>d of
segmental concrete panels (from Mitchell, 1987).
 427

enough geosynthetic tension is mObilized to reach equilibrium. This

is in contrast to steel reinforcement where the reinforcement tension
will be mobilized at comparatively small values of horizontal
extension (i.e., €h = 0.01 to 0.1%).
Geosynthetic tensile stiffnesses are much lower than steel tensile
stiffnesses (i.e., Ss .. 25 to 600 5g). Thi s fact was used as
justification for uncoupling the stress-strain response of an element
of soil from the tension-strain response of the reinforcement.
Calculation of horizontal force equilibrium for the soil and
reinforcement yielded the strains and stresses in the soil and
reinforcement (Figure 6). Extension of the analysis to full-scale
retaining wall structures was discussed and the influence of
construction procedures were qualitatively considered. It was
concluded that due to their low tensile stiffnesses, geosynthetics
will be less effective than steel reinforcement in suppressing
hori zonta 1 soil strains. As a result 1arger soil shear stresses wi 11
be mobilized in geosynthetic reinforced soil walls than in steel
reinforced walls. The wall fill in geosynthetic reinforced soil walls
will be closer to active plastic conditions. For these reasons, limit
state analysis procedures should be more applicable to geosynthetic
reinforced soil walls than to steel reinforced soil walls. In the
next section of this paper, limit state analysis procedures will be
reviewed.
3. LIMIT STATE ANALYSIS METHODS
Two approaches to limit state analysis are reviewed: (i)
approaches based on force or moment equilibrium of a wedge (or block
of soil with some other shape such as a circle, log-spiral, etc.) with
one side of the wedge coincident with the potential failure surface;
and (ii) approaches based on plasticity solutions.
3.1 Limit Equilibrium Solutions
Approaches based on limit equilibrium calculations all involve the
separation ,of the reinforced soil into two zones (Juran and Schlosser,
1978): an active zone, limited on one side by the potential failure
surface which is also assumed to be the locus of maximum reinforcement
tensions; and a resistant zone which provides reinforcement anchorage.
The stress and strain states within the active and resistant zones are
indeterminate and only the shear and normal stresses on the potential
failure surface can be defined. Limit equilibrium solutions vary
mainly by the shape of the assumed failure surface and by the way that
the reinforcement forces are introduced into the equilibrium
equations. The most commonly assumed shapes for the failure surface
are shown in Figure 11 and include a straight wedge, two-part wedge,
circle, and log-spiral.
The simplest assumption regarding the shape of the potential
failure surface is that it is a straight wedge. This assumption was
proposed by Coulomb in 1776 and has traditionally been used to
determine active pressures against yielding retaining walls with
smooth or rough surfaces. Graphical solutions have been developed
for varying values of wall friction (e.g., Terzaghi and Peck, 1967)
and comparisons with more accurate solutions show that the error
 428

(b) .

POTENTIAL ~-'=\~ REINFORCEMENT

FAILURE
SURFACES

TENSILE
ARC /
X
/
C>( / I--~

PRINCIPALA/!>
TENSIl.E
STRAIN
POTENTIAL .-::==----
FAILURE SURFACES

Figure 10. Idealized zero-extension trajectories (0 and ~

trajectories) and potential fail ure surfaces presented by
Bassett and Last (1978) for: (a) a flexible cantilever
wall bending about its toe (data from Milligan, 1974); (b)
a reinforced soil wall with very stiff (e.g.,metallic)
rei nforcement.
 429

associated with the use of a straight wedge for active pressure

calculations are small. For the simple case of zero wall friction and
horizontal backfill with uniform surcharge, the straight wedge
solution is identical to that obtained with Rankine theory (a
plasticity solution), both predicting a critical wedge passing through
the wall toe at an angle of (45°-$/2) from the vertical. Application
of the straight wedge to reinforced soil consists of introducing
reinforcement forces into the equilibrium equations. Early
theoretical formulations for the analysis of Reinforced Earth walls
used Coulomb wedges (e.g., Schlosser and Vidal, 1969; Lee et a1.,
1973). Implicit in the use of this method is the assumption that the
reinforcing elements have a negligible effect on the stresses and
strains within the soil mass. Subsequent investigations (e.g.,
Bague1 in, 1978; Juran and Schlosser, 1978) have shown that thi s
assumption may not be appropriate for soils reinforced with steel.
The straight wedge failure surface (Figure 11a) is a special case
of the more general two-part wedge failure surface (Figure llb). The
straight wedge can be used to evaluate the critical potential failure
surface for simple loading configurations and for calculating the
reinforcement tension along that surface. However, straight wedges
cannot be used to investigate failure surfaces involving sliding of
the reinforced soil over a layer of reinforcement or failure surfaces
that may be critical for external stabil ity. In addition, fai lure
surfaces observed in small-scale tests, numerical studies, and
reinforcement tensions in stable prototype structures, all suggest
that the critical failure surface in steel reinforced structures is
better modeled with a two-part wedge (or a curved failure surface)
than a straight wedge. Two-part wedge limit equilibrium models have
been proposed by Romstad et a1. (1978) and Jewell (1985b) for
reinforced soil walls, and by Stocker et a1. (1979) for nailed soil
walls. Two-part wedge models have been used for the analysis of
geosynthetic reinforced slopes by Murray (1984), Jewell et a1. (1984),
Schmertmann et a1. (1987), and others.
Circular and logarithmic-spiral failure surfaces (Figures 11c and
lld) were .suggested by Juran and Sch 1osser (1978) in conj unct i on with
the requirement for the failure surface to be vertical at the top of
the wall. This requirement was derived from their observation that
steel reinforcement suppresses horizontal soil strains in the upper
portions of the active zone, causing Ko conditions to prevail. The
kinematic condition resulting from Juran's and Schlosser's
requirement is one of wall rotation about a hinged crest. Juran and
Schlosser hypothesized that failure resulted from shear along a very
thin surface between rigid active and resistant blocks. They found
that both the log spiral and circular failure surfaces resulted in
reasonably good predictions of the critical heights of small-scale
model walls. These predictions were better than those based on
Coulomb or Rankine theory. However, only the log-spiral surface
provided a reasonable estimation of the shape and location of the
failure surface. Terzaghi (1943), and more recently Milligan (1983)
and Fang and Ishibashi (1985), have shown that the lateral pressure
distribution associated with walls having a fixed crest is
significantly different than the triangular distribution associated
with rotation about a hinged toe.
430

( a). (b).

RESISTA NT
ZONE

REINFOR CEMENT

L
-I

(c) . (d) .

Figure 11. Common shapes for potent ial failure surfac es for limit
equilib rium stabil ity analys is: (a) straig ht wedge;. (b)
two-part wedge; (c) circle ; and (d) logarithmic spiral
 431

A logarithmic-spiral failure surface has been proposed by

Leshchinsky and Perry (1987) for the analysis and design of
geosynthetic reinforced soil walls. Leshchinsky and Perry also assume
that, at incipient failure, the geosynthetic deforms at its
intersection with the failure surface so that it is orthogonal to the
radius of the logarithmic-spiral. Gourc et a1. (1986) also consider a
reori enta t ion" of the rei nforcement at its intersect i on wi th the
II

failure surface in their "displacement method" of analysis. Gourc et

a1. assume a circular failure surface separating active and resistant
zones. Using assumptions about the deflected shape of the
geosynthetic in the vicinity of the failure surface, they calculate
reinforcement deformations based on a sOi1-geosynthetic e1asto-p1astic
interface deformation model. An additional feature of Gourc et a1.'s
model is that it allows slip between the soil and reinforcement.
Limit equilibrium analysis models utilizing circular slip surfaces
have also been proposed for geosynthetic reinforced embankments by
Studer and Meier (1986) and Ruegger (1986).
3.2 Plasticity Solutions
Classical plasticity methods can be used to develop ultimate loads
for simple rigid-plastic soils under plane strain conditions.
Reinforcement tensions required for equilibrium can be deduced from
the results. Stresses and strains outside of the plastic zone are
indeterminate. Rigorous plasticity solutions have only been developed
for a few simple cases due to their complexity. Approximate solutions
using lower bound stress fields and upper bound kinematic analyses
extend the range of available solutions. Incremental e1asto-p1astic
numerical formulations incorporating anisotropic, strain-hardening
constitutive relationships vastly increase computational
possibilities. However, the complexity of these formulations and the
time involved in using them preclude their use in all but research
applications.
Rankine's solution for a cohesionless, rigid-plastic soil
represents the simplest state of plastic equilibrium for a retaining
wall. S01utions for the more complex cases of walls with surface
friction and sloping backfills were solved using lower bound stress
fields by Caquot and Kerisel (1949) and Soko10vski (1965). Meyerhof
(1980) provides a brief overview of recent work related to the use of
plasticity theories to determine failure mechanisms and stress fields
behind retaining walls with various boundary stress and deformation
conditions, including walls rotating around a hinged top. Scott
(1985) recently presented a comprehensive overview of the history of
the use of plasticity solutions in soil mechanics.
3.3 Comparison of Solutions
Comparisons of several different limit state analysis procedures
were presented by Morgenstern and Eisenstein (1970), Figure 12. These
comparisons are for a retaining wall free to rotate about its toe,
backfilled with a cohesionless, rigid-plastic soil with a horizontal
top surface and various amounts of wall friction. It can be seen that
for active plastic conditions, the various limit equilibrium and
plasticity solutions agree to within about ±10% for equal values of
wall friction. For the case of zero wall friction, the predicted
432

(a).
1201 1201 r - - - - . - - - , - - - ,
~ ~
0 o
S S
I----t---t---::r:>'l Rankine ~ 100 to f--..- r..-::..-:...-.-i.:...--r-::::-o--1 Ro nkin e
>-
a o ~F::;:;';:':~
....."
..c:
801
"
..c:
..... 80l:.

"c ~
'c"
c
~
c
0 601 o 60:1;
:'5.
u.J
- - Rank ine '"
UJ
. -.. Coulomb
0 401 ~
«
.....
- - - - Sokolovski
.....
40%

zu.J -·-··lanbu
.-.•.•.. Brinch Honsen
Z
UJ
U
201 _ .. _. log Spira I ~ 201
'"w
"-
UJ
"-
OX
10' 20' 30' 40' 011':-0.--:-::'20:::-.--:3:'::07".--:40'

if> if>

(b). s •.f..
120% r - - - - . - - - , - - - , 120% r-_--._2--,r-_-,

g ~

;
1001. Rankine o 100% Rankine

"~ ~~-:'-~:
~ 80'l. r-:-::~~i'~""'"
~. ·-·-t~
g
..c:
80Y.
.....
~" 60% 1" WI.
c
o __ Rankine o
'" ·-·Coulomb '"
u.J 401', u.J 40%
- - Sokolovski
o 0
«
.....
- - - Jnnbu «
.....
~ 20%
------ Brinch Honsen
zu.J 20%
- " - log Spiral
u U
"""-
u.J '"
UJ
"-
0"'1'':-0':---:20:::'--==30::-'--:40' 01'.
10' 20' 30' 40'
if>

Figure 12. Comparisons of active pressures calculated using various

limit equilibrium and plasticity analysis methods for a
flexible retaining wall yielding about its toe: (a) total
active earth pressures for various theories; and (b)
horizontal components of total active earth pressures
(from Morgenste rn and Ei sens te in, 1970). S represents
the angle of friction between the retaining wall and soil.
 433

failure surface and horizontal stresses required for equilibrium are

those given by Rankine's solution. From these results, it is
concluded that if the boundary conditions (surcharges and wall
friction) are properly accounted for, the selected limit equilibrium
or plasticity analysis method will not have a large effect on the
computed equilibrium reinforcement tensions (unless the considered
fa il ure surfaces are constrained as is the case with Juran and
Schlosser's logarithmic spiral procedure). Therefore, in the
remainder of this paper, calculations will be based on a simple two-
part wedge limit equilibrium model.
4. ANALYSIS OF GEOSYNTHETIC-REINFORCED SOIL WALLS
In this section, geosynthetic-reinforced soil walls are analyzed
using the two-part wedge model shown in Figure 13. The walls are
assumed to rest on competent foundations. The influence of the facing
unit and construction procedures on wall behavior are neglected. The
important variables considered are the strength of the soil, tensile
stiffness of the geosynthetic reinforcement, and wall geometry. The
two-part wedge analysis will be used to investigate geosynthetic
forces and strains under working conditions. This technique for
evaluating reinforcement tensions was used by Jewell (1985b). In
general, the application of limit equilibrium procedures to working
conditions is questionable. The authors suggest however, that these
procedures may be applicable to the extent that a plastic zone
develops in the geosynthetic-reinforced fill under working conditions.
An IBM PC computer program was written to analyze two-part wedges
with failure surfaces emerging behind the wall crest and at pOints up
the front face of the wall. The program, written in nondimensiona1
form, models inters1ice friction, wall face friction and surcharge
loads. The program automatically analyzes, a very large number of
potential failure surfaces and calculates the required reinforcement
force to maintain equilibrium on each surface. The soil is assumed to
be a uniform frictional material complying with the Mohr-Coulomb
failure criterion. Two-part wedge failure surfaces involving
horizontal ,sl iding of soil over reinforcement account for reduced
shear strength along the sOil-reinforcement interface through use of a
coefficient of interaction, ll, defined as the ratio of the soil-
rei nforcement interface shear strength to the soil shear strength. A
value of II = 1.0 was used for the analysis presented herein. The
interwedge friction angle, 1jI, was found to have only a very minor
influence on the analysis results (1jI = 0 was used herein).
4.1 Global vs. Local Analysis
Jones (1985) has pOinted out that internal stability calculations
for rei nforced soil wa 11 s fa 11 into one of two categori es: (i) those
in which local stability is considered in the vicinity of every single
strip or element of reinforcement; and (ii) those in which the overall
stability of wedges or blocks are considered. Bolton et al. (1978)
considered Reinforced Earth walls as an assemblage of individual
anchors, with each anchor holding back a small area of facing. They
concluded that local equilibrium must be evaluated, since, if anyone
layer of reinforcement breaks, the other layers will successively
break because of sudden load transfer. Both Bolton et al. and
434

DEFINITIONS:
W= Soil weight
N = Normal force on potential failure surface
S = Shear resistance on potential failure surface
P = Interwedge force
T = Horizontal reinforcement force
Q = Surcharge loads
8 = Angle of potential failure surfaces
~ = Angle of surcharge loads
~ = Angle of tnterwedge force

Figure 13. Two-part wedge model used for reinforced soil wall
stability analysis.
 435

Segrestin (1979) point out that the use of a wedge analysis assumes
that all reinforcing elements crossing the failure surface yield
simultaneously. They note that calculated factors of safety based on
Coulomb wedge calculations are larger than those based on local
equil ibrium of individual reinforcing elements. As a result, some
widely used design procedures (e.g., British Dept. of Transport
(1978» incorporate both local and global stability calculations.
Stri ct adherence to 1oca 1 stabi 1 ity requi rements may be overly
restrictive for geosynthetics because: (i) they are much more ductile
than steel; and (ii) they occupy a substantially larger volume in the
reinforced zone (they are in contact with many more soil particles)
than steel. Geosynthetics pick up increasing tensile loads with
increasing tensile strains up to and past the point at which the soil
is yielding. It is reasoned that if local overstressing occurs durjng
the loading process, the soil and reinforcement will deform locally,
increasing the tensile force carried by the reinforcement. Soil shear
stresses generated as a result of differential soil strains should
result in load transfer between reinforcement layers. The relatively
large volume of geosynthetic reinforcement should enhance load
transfer.
Strict adherence to local stability requirements may not be
necessary with geosynthetic reinforcement. However, local stability
can be easily achieved by doing a thorough analysis that assesses the
required reinforcing forces on a wide range of potential failure
surfaces, and then ensuri ng that the rei nforcement di stri buti on
provides the required forces on all surfaces. The procedure described
below for determining the reinforcement force distribution satisfies
local stability requirements.

4.2 Reinforcement Force Distribution

In keeping with the first part of this paper, it is assumed that
the stress-strain response of the soil can be uncoupled from the
tension-st~ain response of the geosynthetic and that locally €h • ft.
With these assumptions, the two-part wedge model shown in Figure 13
can be used to estimate the distribution of reinforcement tensions
required to maintain wedge equilibrium. For any considered wedge, the
total required horizontal reinforcement force is assumed to have two
components: a resultant due to the weight of the soil fill acting at
the one-third height of the wedge (i.e., resulting from a triangular
distribution of required reinforcement tensions); and a resultant due
to surcharge loads acting at the one-half height of the wedge (i.e.,
resulting from a uniform distribution of required reinforcement
tensions).
In order to investigate the distribution of reinforcement tension
with horizontal (x) and vertical (z) distance from the wall crest the
following multi-step analysis was carried out:
(i) The geometry of the soil fill beh i nd the wa 11 face was
nondimensionalized by dividing the vertical and horizontal
axes by the wall height. The fill was then subdivided
into square elements, the side length of each being H/20.
 436

~ X/H 0.5 1.0 1.5

--r----.--- 0
Z/H

H 0.5

POTENTIAL
8LIP SURFACE

klJ '" Kr ~Z/H): - (Z/H):] EL!MENT (lJ)

\ ~_IZ;"b
~-(Z/H).
+ K. ~Z/H)b -(Z/H).]

SLIP SURFACE

Figure 14. Interpretation of two-part wedge analysis results to

estimate the distribution of required reinforcement forces
per unit area, Tij (kPa), in reinforced soil walls.
 437

Each soil element was assigned to an element of a 20 by 30

matrix (Figure 14).
(ii) A trial wedge was selected and equilibrium calculations
were carried out to determine the normalized total
reinforcement force, Kr , due to the weight of the soil
fill alone, and the normalized reinforcement force, Ks '
due to surcharge loads and zero fill weight. Trial wedges
with their toes at all levels of the facing were
considered.
(iii) Kr and Ks for each trial wedge were apportioned to the
elements of the matrix intersecting the considered wedge
(Figure 14); the portion of Kr apportioned to each element
increased proportionally with the depth of the element
be low the crest and wi th the port i on of the potent i a 1
failure surface intercepted by the element (as shown by
the equation given in Figure 14). The portion of Ks
apporti oned to each element increased proporti ona lly with
the portion of the potential failure surface intercepted
by the element.
(iv) The above steps were repeated for a very wide variety of
wedges (several thousand wedges for each set of the input
variables, $, W, and q/yH); the maximum normalized tensile
force for each element, kij' was stored in the matrix.
When the search was completed, the array of maximum kij
values was output.
(v) The distribution of localized reinforcement forces per
unit area (normal to the wall face), Tij (kPa) required to
maintain equilibrium at the location of any matrix element
was determined by using the kij matrix in the equation:
(13)
• where y (kN/m3) is the unit weight of the wall fill and H
is the wall height.
The above procedure for estimating Tij is conservative (i.e, gives
a high value of Tij) because it implicitly assumes that the localized
reinforcement forces are mobilized everywhere at the same time (i.e.,
the soil shear strength is simultaneously mobilized on all of the
trial wedges).
Figure 15 shows the reinforcement force distribution obtained for
the case H = 10 m, $ = 35°, y = 20 kN/m 3 and q = O. The
distinguishing features of the Tij distribution include:
reinforcement tension per unit area that increases 1 inearly with
depth, but is constant along hori zonta 1 planes 1oca ted between the
wa 11 face and ali ne at an ang 1e of 45° - $/2 from the wa 11 face
(active Rankine state); decreasing reinforcement tension with
increasing distance behind the active zone; and a zero force line
extending from the base of the wall back into the fill. This plot is
analogous to one presented previously by Jewell (1985b). An alternate
way to illustrate the required force distribution is through contour
438

2.0 4.0 8.0 0.0 10.0 12.0 1• • 0

I I I I I I I I

T
Z(m) /
/
/

2.0 / ,Z ZERO FORCE LIN

/ ,/

--
! , / !ifILL PROPERT lEa
Til
i .',. 3So
/ ./ 'trw 2 2,0 kN/ra

----
I

./

-----
0 .0
I

' / T lI(kP.'
I

. / ~:o
~
0.0 / 30 .. 0
50 eo
./
~~
10.0
/ //

Figure 15. Distribution of required reinforcement forces per unit

area, Tij (kPa), required to maintain equilibrium in a
10 m (33 ft) high reinforced soil wall.
 439

plots such as those shown in Figure 16. These plots show all the same
features as Figure 15. In addition, they allow rapid evaluation of
Tij anywhere in the fill. The effect on Tij of soil strength ($) and
un1form vertical surcharges is shown in F1gure 16.
Figures 15 and 16 show ~that to satisfy local equilibrium
requirements, tensile forces are required to a distance back from the
wall crest equal to 1.0 to 1.5 times the wall height (X/H = 1.0 to
1.5). In practice, reinforcement lengths are typically on the order
of L/H = 0.8. The wedge of unreinforced soil between the reinforced
zone and zero force line exerts a thrust and overturning moment on the
reinforced soil mass. Calculations show that for L/H = 0.8, $ > 35°,
and q/yH < 0.2, the overtu rni ng moment due to the th rust of the
retained fill is insignificant, as is the calculated vertical stress
increase within the reinforced soil mass due to the overturning
moment. Therefore, calculation of vertical stresses using the
"trapezoidal" or "Meyerhof" distributions is unnecessary. Further,
Jewell (1985b) noted that reinforcement truncation short of the zero
force line causes the locus of maximum reinforcement tensions to move
toward the wall face. While this effect may be significant for steel
strip reinforcement, the effect was found to be insignificant for
geosynthetic reinforcement (due to the short required bond lengths for
geosynthetics compared to steel strips).
4.3 Reinforcement Defonmations
Reinforcement tensile strains, €ij' were calculated from the local
reinforcement tensions as follows:

(14)
where Tij (kN/m) was obtained from Equation 13 and 5g was calculated
as shown in Figure 4. In Equation 14, Tij and 5g are independent
variables (i.e, the reinforcement tensile st1ffness does not influence
the state of stress). The total reinforcement elongation at any
elevation o~(z) was obtained from:

(15)

where L (m) is the length of reinforcement.

Figure 17 presents €max and or values for a reinforced soil wall
with H = 10 m, $ = 35°, Ym = 20 kN/m 3 , and q = O. Results are
presented for a range of 5g values and for 5s = 70,000 kPa. 5g and 5s
are assumed to be constant with depth (uniform reinforcement layout).
For these conditions, the maximum calculated reinforcement strain
occurs at the bottom of the wall and the maximum calculated
reinforcement elonga~ion occurs at (z/H) - O~~. For 59 = 2,000 kPa,
or max = 80 mm (3.2 1n) and 0r,ma,JH = 8 x 10 • In contrast, for 5s
= 70,000 kPa, Or max/H = 4 x 10 I . The form of these results are
cons i stent with 'those reported by Mi tche 11 (1987) and shown in
Figure 9.
440

0
~

g %

~
" ~

"
w
z
~

w
"I
:
-<
o
:
_If
0
II
O .. I.00.,!...
r"
!:
0

~ "a:
~

~
~

0
~

"
0

o
~ ~-------,-------,--------r-------,

" t----T---j--
o

"Io~ • "
:.11
III

o
N 0

-O-c..g
%
)co

13

o
~

"t---r~r-Ti-r~--t-ir
o

..
~ o ~

"o
,0 ~ ~

Figure 16. Contours of Reinforcement force per unit area, (T . .IyH),

required to maintain equilibrium in a reinforced ~5il wall.
Contours obtain using two-part wedge analysis results.
 441

~6Lr~(m~m~)~7+5~____~5~OL-____~2~5~____~r- ______~2______~4 £max (~)

Fill PROPERTIES
.p ' : 35 0
3
~m = 20kN/m

Sa: 70,000 kPa

Figure 17. Reinforcement elongation (&r) and maximum tensile strain

required to maintain equilibrium in a reinforced soil
wall.
 442

SgCkPal
100 1,000

FILL PROPERTIES
fI'= 35° 2,000

75 ~m = 20kN/m 3

3,000
50

4,000

8,000
211

J S.= 70,000 kPa

o ---- --- H(m)
8.0 8.0 10.0

2,000

6r • 150mm

4,000

Figure 18. Relationship between wall height (H), reinforcement

stiffness per unit soil area (Sg and Ss), and
reinforcement elongation (or) in a reinforced soil wall.
 443

Figure 18 shows the relationship between wall height, H,

reinforcement stiffness per unit area (Sg and Ss), and reinforcement
elongation (or). For a given allowable reinforcement elongation, the
maximum wall height is dependent on the reinforcement stiffness. It
is important to note that or is not the total wall face displacement,
as measured from a vertical plane through the wall toe at the start of
construction. The total displacement will increase with distance
above the wall base (neglecting initial wall batter). The total
displacement can only be obtained through an analysis that takes into
account the incremental nature of wall construction. or should be
interpreted strictly as the reinforcement elongation at any level.
Further, or will be affected by the boundary conditions imposed on the
soil and reinforcement.

5. DESIGN OF GEOSYNTHETIC REINFORCES SOIL WALL

5.1 Method of Analysis
The analysis procedures described in Section 4 can be used to
design geosynthetic-reinforced soil walls. The design process must
take into account all potential failure mechanisms through, under, and
behind the reinforced soil mass. Various shapes for the potential
failure surface can be assumed (Figure 11). Two-part wedge failure
surfaces (Figure llb) may offer the best combination of accuracy,
versatility and simplicity for widespread application as a design
tool. Direct sliding of soil on reinforcement can be modeled using
two-part wedges with a horizontal lower failure surface. The soil
shear strength along this lower surface is reduced by the soil-
reinforcement coefficient of interaction, p, obtained from the results
of direct shear tests. Two-part wedges can be used to investigate
global stability. The design process should also include an analysis
for adequate foundation bearing capacity.
In most geosynthetic-reinforced soil walls, the length of
reinforcement is controlled by external stability considerations
(resistance, to direct sliding, bearing capacity, or global stability)
rather than by reinforcement pullout from the resistant zone behind
the potential failure surface. Due to the comparatively high
geosynthetic reinforcement volume (0.2 to 0.5% of the reinforced fill
volume), stress levels in the reinforcement are low compared to those
in stee 1. Th is fact, coup 1ed with geosynthet i cs 1arge embedded
I

surface areas, and their generally good bond with granular soils,
preclude concern over the length required to prevent pullout for all
but the uppermost layer(s) of reinforcement. With geosynthetics,
there is more concern over reinforcement rupture than pullout.
5.2 5011 and Reinforcement Properties for Design
The · design process necessitates selection of soil and
reinforcement properties for design and the selection of appropriate
factors of safety.
For des i gn, soil strengths and geosynthet i c tens ions shou 1d be
selected at compatible values of strain. Since geosynthetics are
extensible and ductile, it appears both logical and conservative to
444

define a limit state for design corresponding to mobilization of the

large-strain, constant-volume strength of the soil. Recommendations
for selection of reinforcement tensions compatible with the use of $cv
were first discussed by McGown et al. (1984a,b) and subsequently by
Jewell (1985) and Bonaparte et al. (1985, 1987) and are not repeated
here.
The selection of soil properties to determine working strains and
deformations is more complex due to progressive soil strength
mobilization (e.g., Murray, 1987) and progressive strain softening
(e.g., Rowe, 1969b). Practically, the use of $p is acceptable for
analyses of working conditions, as experience with existing
geosynthetic reinforced soil walls indicates relatively small
reinforcement strains under these conditions.
In most practical applications, both $cv and $p can be 'based on
plane strain deformation conditions. Jewell and Wroth (1987) have
recently presented a comprehensive review of the relationship between
plane strain ($ps) and direct shear angles of friction ($ds). They
make recommendations for the determination of $ps from direct shear
test results.
5.3 Factors of Safety
Design of reinforced soil walls involves the application of an
overall factor of safety against reinforcement rupture. There are two
ways to incorporate the factor of safety (FS) and both are commonly
used: factor the shear strength of the soil by FS and use thi s
factored strength in design calculations; or, use the actual shear
strength in design calculations and increase the calculated
reinforcement tension by FS. These two procedures give roughly the
same reinforcement design tension (ad) at $ = 35°. For $ = 25°, the
first procedure results in ad about 10% smaller than the value of ad
obtained with the second procedure. At $ = 45°, the reverse is true.
At more extreme values of $, the resul ts based on the two methods
diverge significantly. The design engineer should be aware of this
difference, particularly when dealing with unusual problems or
comparing different designs for the same application (i.e., designs
from different proprietary wall suppliers).
5.4 Design Details
Design of geosynthetic-reinforced soil walls includes many details
that can significantly influence wall behavior. Most important of
these are the type of facing unit and the construction procedures. As
indicated in Section 2.5, the type of facing unit and construction
procedures strongly influence the wall failure mechanism and soil
strain field.
Design of the reinforcement-to-panel connection is important not
onlj with respect to the connection strength, but also with respect to
deformation of the wall facing. For instance, Berg et al. (1987)
attributed 6 to 9 mm of measured face movement in a 5.3 m (17.4 ft)
high geogrid-reinforced soil wall to insufficient pretensioning of the
geogrid-to-panel connection. Compaction may result in movement of the
wall facing as well as compaction-induced soil stresses and
 445

reinforcement tensions (Coll in, 1986; Mitchell, 1987). Finally,

foundation preparations and drainage provisions wi 11 both affect wall
behavior.

6. ANALYSIS OF RMC REINFORCED SOIL WALLS

In this section, the analysis procedures described in Section 4
are applied to two geogrid-reinforced soil walls constructed and
loaded in 1986/1987 at the Dol phi n Structu res Laboratory, -Roya 1
Military College (RMC) , Kingston, Ontario. For completeness, a brief
description of the RMC test walls is included. A more thorough
discussion of the materials, design, and construction of the walls can
be found in other papers presented at the NATO Workshop.
6.1 Description
Cross-sections and front views of the two RMC reinforced soil
walls are shown in Figure 19. Both walls were 3.0 m (9.8 ft) high and
were constructed with medium dense (relative density" 55%) coarse
sand and four layers of 3.0 m (9.8 ft) long TENSAR SR2 geogrid
reinforcement. Geogrid strips were fastened to the wall facing panels
prior to panel erection. Each geogrid layer was pretensioned to 0.4
kN/m (27 lbs/ft) to remove slack prior to backfilling over the layer.
The fill was compacted using a small vibrating plate tamper. One wall
was constructed with incremental timber facing panels. The 0.75-m
(2.5-ft) high panels were separated by horizontal layers of foam
rubber filler. The bottom-most course of panels rested directly on a
concrete leveling pad that was restrained from forward movement. A
second wall was constructed using a propped vertical timber facing
which also rested on a horizontally restrained leveling pad. Three
levels of props prevented movement of the timber facing during fill
placement and compaction. The props were removed when the fill height
reached the top of the wall. A uniform vertical surcharge of 12 kPa
(250 psf) was placed on top of each of the walls immediately after
construction. The surcharge was achieved by placing a 750 mm (2.5 ft)
layer of uncompacted loose sand over the entire area of sand backfill.
Additional vertical surcharges were applied using pneumatically
inflated neoprene bags.
6.2 Selection of Reinforcement and Soil Parameters
The rei nforcement input parameters are the tens i 1e stiffness, J g
(kN/m), and the tensile force per unit width at failure, af (kN/m).
Both J g and af are strain-rate and temperature dependent. Normally,
information on soil-reinforcement bond and reinforcement pullout is
also required. Based on the work of Sarsby (1985), the coefficient of
soil-reinforcement interaction, ~, should be on the order of 0.9 or
more. Simple calculations quickly demonstrate that the reinforcement
1engths a're suffi ci ent to prevent rei nforcement pullout pri or to
rupture.
The required soil input parameters are the plane-strain friction
angle, $p's' and the unit weight y (kN/m"). The dry density (p) of the
sand fill at Dr = 55% was reported to be about 1.79 Mg/m" and the
compaction moisture content was about 3%; therefore, the moist unit
446

SAND/AIR PRESSURE BAG SURCHARG7

(a) .
UNREINFORCED
SAND BACKFILL
-7 SR2 GEOGRID
REINFORCED sOIy
SURCHARGEA
PANELS

---jC--------ti --- - II \ -
I-- -SR2 LAY~ 4 ' ::::::::270:
::::::
Te
o
SR2 LAYER 3

SR2 LAYER 2
::::: :: :::: :: y: (

k:: :::'/'::'"" ::,:"" ":::: ::: < O:i e I

SR2 LAYER I -----Q ': <:::: '0: :: ( :J
- - - -"- - - - L''''.L''''.':;:;·.:i>~,··J;.'4
I

2.4 ...
SAND/AIR PRESSURE BAG SURCHARGE

SURCHARGE
"1\
I\
PANELS
(b).
- j r--

SR2

SR2
T....
U)

"",
~:
SR2 eo

1
oz

SR2

I
• STRAIN GAUGE
_ PRESSURE CELL

CROSS - SECTION VIEW FRONT VIEW

Figure 19. Cross-sections and front views of RMC reinforced soi 1

walls: (a) incremental timber facing panels; (b) propped
vertical timber facing panels.
 447

weight of the fill was about 18.1 kN/m 3 • The sand shear strength was
measured in 60 mm square di rect shear tests and in conso1 idated-
undrained triaxial tests. Direct shear tests carried out on fill
specimens with an initial dry density of Po = 1.78 Mg/m 3 gave a peak
friction angle ($p)ds = 43 0 • Direct shear tests also gave a value for
the large-strain, constant-volume friction angle of ($cv)ds = 40 0 •
Using the equation by Rowe (1969a) relating plane strain and direct
shear friction angles:

tan$ps = (tan$ds) I cos($cv)ds (16)

yields ($p)ps = 510 for $ds = 43° and ($cv)ps = 48° for $ds = 400~
Therefore, $ = 48° represents a lower bound of $ and $ = 51 °
represents an upper bound. The authors have elected to use $ = 45 0
and $ = 500 as approximate lower and upper bounds i n subs~quent
calculations.
The properties of the high density polyethylene (HOPE) ge09rid in
the RMC test wall have been described by McGown et a1. (1984a).
Specific test data on the reinforcement materials used to build the
test walls were provided by the workshop organizers. No temperature
corrections to the test data are required since both geogrid physical
property testing and test wall construction were at 20°C (68°F).
The following secant tensile stiffness values (J g in kN/m) have
been selected by the authors from the results of constant - load creep
tests provided by the workshop organizers:

Load duration (hr)

Tensile Strain (%) 100 1000
0-2 600 500
2 - 5 500 400

For purposes of calculating the critical wall height and factors

of safety, the reinforcement tensile force per unit width at failure
is requ ired. The authors have chosen to se 1ect af based on a 10%
limiting strain criterion. This results in af = 35.3 kN/m for a load
duration of 100 hours and af = 34.0 kN/m for a load duration of 1000
hours.
6.3 Analysis Results
Analysis of the RMC reinforced soil walls was carried out using
the two-part wedge computer program described in Section 4.
Unfortunately, it is not possible to use the two-part wedge model to
differentiate the behavior of the incremental panel and propped panel
walls. This is a significant limitation of limit equilibrium models.
 448

r
------~----_.-------r--·----.I------,-----_.I X(m)
o 2 3
o
-- --- ------ 8
GEOGRID
LAYER
-----
____________
- _
:~ «r (kN/m)
----------- -
4-- r-------------------------------~~--------

1.0

-----
3--r------------~~~-------

-----=~
--- -
FILL PROPERTIES
¢"~45°
lI'm= 18.1kN/m 3

:~
o
2.0 -----
2-r--------~~~~--S-U-R-C-H-A-R~G~E~(k~p~a~)-
----- 0
-----12
--- - -50

1- r------"o::...lo..--------------

Z(m)
3.0

Figure 20. Reinforcement forces per unit width, a r (kN/m) in geogrid

layers. Results for RMC reinforced soil walls obtained
using two-part wedge analysis.
 449

Frictional forces on the sidewalls of the reinforced soil mass were

also neglected.
Results of the analysis are shown in Figures 20 and 21. Figure 20
shows the distribution of required reinforcement forces in each of the
four layers of geogrid reinforcement for $ = 45°. The calculated
maximum reinforcement tensions, ar,max (kN/m), in each geogrid layer
are:

q (kPa) q (kPa)
Geogrid
Layer I 12
I 30 I 50
I 12
I 30
I 50
4 I 3.0 I 5.7 I 8.5I 2.3 I 4.5 I 6.7
3 I 4.3 I 6.6 I 9.1 I 3.4 I 5.1 I 7.1
2 I 6.1 I 8.4 I 10.9 I 4.7 I 6.5 I 8.5
1 I 6.5 I 8.5 I 10.5 I 5.0 I 6.6 I 8.2
I $ = 45° I $ = 50°

The analysis results indicate a maximum reinforcement tension,

ar,max, at q = 50 kN/m 2 (1040 psf) on the order of 8 kN/m (550 lbs/ft)
for $ = 50° and 11 kN/m (750 lbs/ft) for $ = 45°. If the factor of
safety (FS) of the reinforced walls is taken as the ratio of
af/ar ,max, ,then FS - 3 for ar ' max = 11 kN/m (750 lbs/ft).
Reinforcement strains are calculated from reinforcement tensions
using Equation 14. The maximum calculated strains at each geogrid
elevation are shown in Figure 21.
For the case of a 100 hour load duration and q = 12 kPa (250 psf),
Emax = 1.1% for $ = 45° and 0.8% for $ = 50°. For q = 50 kPa (1040
psf), Emax = 1.8% for $ = 45° and 1.4% for $ = 50°. Calculated
strains increase 20% as the surcharge durations increase from 100 to
1000 hours. The calculated maximum strain occurs in the lowest
geogrid layer when q = 12 kPa and in the next-to-lowest layer when q =
50 kPa.
Calculated reinforcement elongations are shown in Figure 21. For
q = 0 to 12 kPa (0 to 250 psf), the maximum elongation takes place
around the mid-height of the wall. For q = 30 to 50 kPa (620 to 1040
psf), however, the maximum elongation occurs in the top geogrid layer.
Maximum geogrid elongations for various surcharges and a 100 hour load
duration are:
450

S, (mm)
30 20 10 o

GEOGRID
LAYER
q(kPa) 50 50 30. 12 0 o 30 50 4
'\ ~12 Y Y i 60
\ \ \
\ \ \ r / \ \

'\\ \ j fLO \\

\ \ \ \
\ \.\ 3

\ \ I
, \ \ \ I .\ \ \ \ \

..
\ \\ \ ~ t
,~\ \ \1
2.0
\\\ \)2
\ I I I
\ I
LOAD DURATION (h,) '\ \
\ I I I '
<>- -

__ - __
- -0 100
1000
"-;\\ \ \
\.\,\\~
\ I
U !
I I
b
1
3.0
Z(m)

&,(mm)
30 20 10 o

GEOGRID
LAYER
q(kPa) 50 50 30 12 30 50

\\ \
4
12 111 \60
\ I '\
\ \ I ,

'\\ \ ( o. \ \ \ \ 3

\ \ I

\'\\\\ ,. \\\\r
\ \ \

\ \\ \ I I
, \ \\ I I I
\\\,
LOAD DURATION (hr)

---
I

0----- --0 100 I I I}

1000 \\\~ ~ !!
3.0

Fig'ure 21. Geogrid elongation (&r> and maximum strain (et> for RMC
reinforced soil walls. Results obtained using two-part
wedge analysis.
 451

q (kPa)
I 12 I 30 I 50

$ = 45°
I or,max (1l1Il)1 11
I 21
I 30

I or,max /H I 0.004 I 0.007 I 0.01

$ = 50°
lor ,max (1l1Il) I 7
I 13
I 20

lor max /H I 0.002 I 0.004 I 0.007

6.4 Discussion of Results

The analysis results shown in Figures 20 and 21 appear reasonable,
particularly given the uncertainties inherent in attempting to use
limit equilibrium procedures to predict working strains and
deformations. The analysis results should only correlate with the
actual behavior of the RMC walls if the actual soil and reinforcement
properties and applied boundary conditions match those used in the
analysis. The properties of the soil and reinforcement used in the
RMC walls are reasonably well defined. Most of the uncertainty rests
with the boundary conditions.
The bounda ry cond it i on on top of the wa 11 s was mode 1ed as a
uniform vertical surcharge. This model is incomplete because it
neglects the fact that the actual surcharge conSisted of 0.75 m (2.5
ft) of loose sand. The sand strength was neglected in the model as
were shear stresses developed at the interface between the top of the
wall and the loose sand surcharge. The net effect of neglecting these
factors is unknown. Plane strain conditions were assumed in the
model, requiring frictionless sides to the RMC test wall facility.
The actual ~olyethylene-soil interface friction angle along the side
walls will be on the order of 12°. This friction will result in soil
shear stresses that will reduce geogrid strains and wall movements.
If a sidewall friction angle of 12° were incorporated into the
analysis, average geogrid strains and tensions would be reduced by
about 10%.
Two different sets of boundary conditions must be considered for
the front of the wall: those associated with the incremental timber
facing panels and those associated with the propped vertical timber
facing panels.
With respect to the incremental panel wall, each panel is attached
to only one geogrid layer. This causes the panels to be statically
unstable unless sufficient normal and shear forces develop between
panels (and between the lowest panel and the leveling pad) to prevent
panel rotation. Either individual panels will rotate significantly
out of alignment or large shear and normal forces will be transmitted
between panels. In this latter case, the foam rubber filler will
452

compress. The filler may become ineffective in minimizing shear

stresses at the soi 1-t imber interface. If soi 1-timber interface
friction is fully mobil ized, lateral earth pressures would be reduced
by about 10%. according to Figure 12. Reinforcement elongations and
strains would also be reduced. Interface shear stresses between the
timber and reinforcement may also reduce 0v in the vicinity of the
wall face to values less than (yz + q).
For the tilt-up panel wall, reinforcement strains in the vicinity
of the wa 11 face wi 11 be contro 11 ed by the amount of wa 11 rota t ion.
This may lead to a geogrid strain distribution near the wall face
different than that predicted from the two-part wedge solutions.
Frictional forces should develop between the tilt-up panels and the
reinforced soil resulting in reduced horizontal soil stresses and
geogrid strains. The frictional forces, coupled with the deformation
constraints of the tilt-up panel, may result in significant deviations
from the conditions assumed in the two-part wedge model.
In conclusion, the authors believe that the boundary conditions
imposed on the RMC reinforced soil walls may significantly influence
the measured soil and reinforcement stresses and strains, particularly
in the vi ci ni ty of the faci ng uni ts. The i nf1 uence of these boundary
conditions cannot be adequately accounted for using the two-part wedge
ana 1ys is. I n general, however, the imposed boundary conditions shou1 d
reduce geogrid strains and elongations below the values calculated
using the two-part wedge model, possibly by as much as 20 to 30%.
7. CONCLUSIONS
An analysis of the influence of reinforcement tensile stiffness on
the behavi or of an element of soil was carri ed out. The ana 1ys is
indicated that the tensile stiffnesses of steel reinforcement are
sufficient to suppress horizontal soil strains and thereby prevent
full mobilization of the shear strength of the soil. As a result, the
calculated lateral stresses in a reinforced soil mass are close to
those associated with an at-rest (Ko) state of stress. The tensile
stiffnesses of polymer materials used in geosynthetic-reinforced soil
walls' are much lower than those of steel reinforcement. The
calculation results indicated that geosynthetic reinforcement will not
reach equilibrium until sufficient strain has occurred to fully
mobilize the shear strength of the soil. As a consequence, an active
(Ka) plastic state of stress should develop in geosynthetic-reinforced
soil walls.
It was suggested that classical limit state analysis procedures,
modified to take into account the presence of reinforcement forces,
are appropriate for analysis and design of geosynthetic-reinforced
soil walls. A limitation of these procedures as they apply to
reinforced soil is that they involve an uncoupling of the soil
stresses and reinforcement forces. It was argued that this uncoupling
is more acceptable for extensible geosynthetic reinforcement than for
inextensib1e steel reinforcement. Available limit equilibrium and
plasticity analysis methods were reviewed and it was noted that the
selected method of analysis has little influence on the calculated
required tensile forces for active earth pressure problems.
 453

The parametric analyses used in the latter part of paper were

based on a two-part wedge limit equilibrium model. The model is
conservative and results in reinforcement tensions consistent with an
active (Ka) Rankine state of stress. The model was used to investigate
the influences of reinforcement tensile stiffness, soil shear strength
and wall geometry on the distribution of reinforcement tensile
strains, forces and elongations in reinforced soil walls. The analysis
results show reinforcement strains and elongations inversely
proportional to reinforcement stiffness. Conversely, for a given
reinforcement stiffness, calculated reinforcement strains and
elongations increase rapidly with increasing wall height and
decreasing soil strength. It was suggested that for geosynthetic
reinforcement, overturning moments and vertical pressures due to the
thrust of the retained fill can be neglected. Therefore, it is
unnecessary to incorporate the "Meyerhof" or "trapezoidal" vertic;a1
stress distributions into design calculations. Vertical stresses
should be based on the fill weight and the applied surcharge loads.
In the last part of the paper, the RMC reinforced soil walls were
analyzed using the two-part wedge model. It was noted that 1imit
equilibrium models cannot be used to differentiate walls built using
different facing units and construction procedures. Results from the
analysis were presented. The results indicate a maximum reinforcement
strain on the order of 1% for a surcharge load of 12 kPa (250 psf) and
1.5% for a surcharge load of 50 kPa (1040 psf). A review was carried
out to assess whether the boundary conditions used in the 1imit
equilibrium model matched those in the RMC walls. It was concluded
that the facing units used for construction of both walls impose
significant boundary stresses on the front of the reinforced soil
mass. Boundary stresses also result from the dead load surcharge on
top of the reinforced soil and from friction along the sides of the
test facil ity. These boundary stresses should reduce geogrid strains
and elongations below the values calculated using the two-part wedge
model, possibly by as much as 20 to 30%.
454

REFERENCES
Bacguelin, F. (1978), "Construction and Instrumentation of Reinforced
Earth Walls in French Highway Administration", Proceedings, Symposium
on Earth Reinforcement, American Society of Civil Engineers,
Pittsburgh, pp. 186-201.
Bassett, R.H. and Last, N.C. (1978), "Reinforcing Earth Below Footings
and Embankments", Proceedings, Symposium on Earth Reinforcement,
American Society of Civil Engineers, Pittsburgh, pp. 202-231.
Berg, R.B., La Rochelle, P., Bonaparte, R. and Tanguay, L. (1987),
"Gaspe Peninsula Reinforced Soil Seawall-Case History", Proceedings,
Symposium on Soil Improvement, ASCE Geotechnical Special Publication
No. 12, Atlantic City, pp. 309-328.
Berg, R.R., Bonaparte, R., Anderson, R.A., and Chouery, V.E. (1986),
"Design, Construction and Performance of Two Geogrid Reinforced Soil
Retaining Walls", Proceedings, Third International Conference on
Geotextiles, Vienna, Vol. 2, pp. 401-406.
Bolton, M.D., Choudhury, S.P. and Pang, P.L.R. (1978), "Reinforced
Earth Walls:-A Centrifugal Model Study", Proceedings, Symposium on
Earth Reinforcement, American SOCiety of Civil Engineers, Pittsburgh,
pp. 252-281.
Bonaparte, R. and Berg, R.R. (1987), "Long-Term Allowable Tension for
Geosynthet i c · Rei nforcement", Proceedi ngs, Geosynthet i cs '87, New
Orleans, Vol. I, pp. 181-192.
Bonaparte, R., Holtz, R.D. and Giroud, J.P. (1985), "Soil
Reinforcement Design Using Geotextiles and Geogrids", Geotextile
Testing and the Design Engineer, American Society for Testing and
Materials, Philadelphia, pp. 69-115.
British Dept. of Transport (1978), "Reinforced Earth Retaining Walls
and Bridge Abutments for Embankments", Technical Memo BE3/78.
Caquot, A. and Kerisel, J. (1956), "Traite de Mecanique des Sols",
Gauthier-Villars, Paris.
Collin, J.G. (1986), Earth Wall DeSign, Ph.D. Dissertation, University
of California, Berkeley, 440 p.
Duncan, J.M. and Chang, C.Y (1970) "Nonlinear Analysis of Stress and
Strain in Soils", Journal of the Soil Mechanics and Foundations
Division, ASCE, Vol. 96, No. SM5, Sep, 1629-1653.
Fang, Y.S. and Ishibashi, I. (1986), "Static Earth Pressures with
Va~ious Wall Movements", Journal of Geotechnical Engineering, Vol.
112, No.3, Mar, pp. 317-333.
Gourc, J.P., Ratel, A. and Delmas, P. (1986), "Design of Fabric
Retaining Walls: The 'Displacement' Method", Proceedings, Third
International Conference on Geotextiles, Vienna, Vol. 2, pp. 289-294.
 455

Handy, R.L. (1985), "The Arch in Soil Arching", Journal of

Geotechnical Engineering, ASCE, Vol. 111, No.3, Mar, pp. 302-318.
Ingold, T.S. (1982), Reinforced Earth, Thomas Telford Ltd., London,
141 p.
Jaky, J. (1944), "The Coefficient of Earth Pressure at Rest", Journal
of the Society of Hungarian Architects and Engineers, pp. 355-358.
Jewell, R.A. (1985a), "Material Properties for the Design of
Geotextile Reinforced Slopes", Geotextiles and Geomembranes Journal,
Vol. 2, No.2, pp. 83-109.
Jewell, R.A. (1985b), "Limit Equilibrium Analysis of Reinforced Soil
Walls", Proceedings. Eleventh International Conference on Soil
Mechanics and Foundation Engineering, San Francisco, Vol. 2, pp. 1705-
1708.
Jewell, R.A. and Wroth, C.P. (1987), "Direct Shear Tests on Reinforced
Sand", Geotechnique, Vol. 37, No.1, pp. 53-68.
Jewell, R.A., Paine, N. and Woods, R.I. (1984), "Design Methods for
Steep Reinforced Embankments", Proceedings, Symposium on Polymer Grid
Reinforcement in Civil Engineering, The Institution of Civil
Engineers, London, pp. 70-81.
Jones, C.J.F.P. (1985), Earth Reinforcement and Soil Structures,
Butterworths Advanced Series in Geotechnical Engineering, London,
183 p.
Juran, I. and Schlosser, F. (1978), "Theoretical Analysis of Failure
in Reinforced Earth Structures", Proceedings, Symposium on Earth
Reinforcement, American Society of Civil Engineers, Pittsburgh, pp.
528-555.
Kondner, R.L. (1963), "Hyperbol ic Stress-Strain Response: Cohesive
SOilS", Joyrnal of the Soil Mechanics and Foundations Division, ASCE,
Vol. 89, No. SM1, Feb, p. 115.
LCPC-SETRA (1979), "Les Ouvrages en Terre Armee", Recommandations et
Regles de l'Art, 195 p.
Lambe, T.W. and Whitman, R.V. (1968), Soil Mechanics, John Wiley and
Sons, Inc., New York, 553 p.
Lee, K.L., Adams, B.D. and Vagneron, J.J. (1972), "Reinforced Earth
Retaining Walls", Journal of the Soil Mechanics and Foundations
Division, ASCE, Vol. 99, No. 10, Oct, pp. 745-764.
Leshchinsky, D. and Perry, LB. (1987), "A Design Procedure for
Geotextile Reinforced Walls", Proceedings, Geosynthetics '87, New
Orleans, Vol. 1, pp. 95-107.
McGown, A•• Andrawes, K.Z., and Al-Hasani, M.M. (1978), "Effect of
Inclusion Properties on The Behavior of Sand", Geotechnique, Vol. 28,
No.3, pp. 327-346.
456

McGown, A., Andrawes, K.Z., Yeo, K.C. and DuBois, D.O. (1984a), "The
Load-Strain-Time Behavior of Tensar Geogrids", Proceedings, Symposium
on Polymer Grid Reinforcement in Civil Engineering, The Institution of
Civil Engineers, London, pp. 11-17.
McGown, A., Paine, N. and Dubois, D.O. (1984b), "Use of Geogrids in
Limit Equilibrium Analysis", Proceedings, Symposium on Polymer Grid
Reinforcement in Civil Engineering, Institution of Civil Engineers,
London, pp. 31-36.
Meyerhof, G.G. (1953), "The Bearing Capacity of Foundations Under
Ecce nt ric and Inc 1i ned Loa ds" , '--P"=ro""'c:..:e:..:e'7d:....:i-:'-'n~g.::.s~,-=-T.:. .h:. :.,',:-;or,-,d=-.;.I:.:..nt~e:..:r...!n.:..;a;-:t:..:.i-70.!.:.nf-!.a1
Conference on Soil Mechanics and Foundation Engineering, Zurich,
Vol. 1, pp. 440-445
Meyerhof, G.G. (1980), "Limit Equilibrium Plasticity ' in Soil
Mechanics", Proceedings, Symposium on Applications of PlastiCity and
Generalized Stress-Strain in Geotechnical Engineering, ASCE, Florida,
pp. 7-24.
Milligan, G.W.E. (1974), "The Behavior of Rigid and Flexible Retaining
Walls in Sand", Ph.D. Dissertation, University of Cambridge, England.
Mill igan, G.W.E. (1983), "Soil Deformations Near Anchored Sheet Pile
Walls", Geotechnigue, Vol. 33, No.1, pp. 41-55.
Mitchell, J.K. (1987), "Reinforcement for Earthwork Construction and
Ground Stabilization", Theme Lecture, Preprint, 'VIII Pan American
Conference on Soil Mechanics and Foundation Engineering, Cartagena,
Aug.
Morgenstern, N.R. and Eisenstein, Z. (1970), "Methods for Estimating
Lateral Loads and Deformations", Proceedings, Specialty Conference on
Lateral Stresses and Earth Retaining Structures, American Society of
Civil Engineers, Ithaca, pp. 51-102.
Murray, R.T. (1984), "Reinforcement Techniques in Repairing Slope
FailtJres", Proceedings, Symposium on Polymer Grid Reinforcement in
Civil Engineering, London, pp. 47-53.
Murray, R.T. (1987), "Factor of Safety ConSiderations Relating to
Reinforced Soil Structures", Preprint, NATO Advanced Research Workshop
on Polymeric Reinforcement in Soil Retaining Structures.
Romstad, K.M., Al-Yassin, A., Hermann, L.R. and Shen, C.K. (1978),
"Stability Analysis of Reinforced Earth Retaining Structures",
Proceedings, Symposium on Earth Reinforcement, Pittsburgh, pp. 685-
713.

Ro~e, P.W., (1969a), "The Relation Between the Shear Strength of Sands
in Triaxial Compression, Plane Strain and Direct Shear", Geotechnigue,
Vol. 19, No.1, pp. 75-86.
Rowe, P.W. (1969b), "Progressive Failure and Strength of a Sand Mass",
Proceedings, Seventh International Conference on Soil Mechanics and
Foundations Engineering, Mexico City, Vol. 1, pp. 341-349.
 457

Ruegger, R. (1986), "Geotextile Reinforced Structures on Which

Vegetation Can Be Established", Proceedings, Third International
Conference on Geotextiles, Vienna, pp. 453-459.
Sarsby, R.W. (1985), "The Influence of Aperture Size/Particle Size on
the Efficiency of Grid Reinforcement", Proceedings, Second Canadian
Conference on Geotextiles and Geomembranes, Edmonton, pp. 7-12.
Schlosser, F. (1978) liLa Terre Armee, historique development actuel et
Futur", Proceedings, Symposium on Soil Reinforcing and Stabilizing
Techniques, NWSIT/NSW University, pp. 5-28.
Schlosser, F. and Vidal, H. (1969), "La Terre Armee", Bulletin de
Liason du Laboratoire des Ponts et Chaussees, No. 41, Nov, pp. 101-
144.
Schmertmann, G.R., Bonaparte, R., Chouery, V.C. and Johnson, R.
(1987), "Design Charts for Geogrid Reinforced Soil Slopes",
Proceedings, Geosynthetics '87, New Orleans, Vol. 1, pp. 108-120.
Scott, R.F. (1985), "Plasticity and Constitutive Relations in Soil
Mechanics", Journal of Geotechnical Engineering, ASCE, Vol. 111,
No.5, May, pp. 563-605.
Segrest in, P. (1979), "Des ign of Rei nforced Earth Structures Assumi ng
Failure Wedges", Proceedings, International Conference on Soil
Reinforcement, Paris, Vol. 2, pp. 163-168.
Sokolovski, V.V. (1965), Statics of Granular Materials, Pergamon
Press, London.
Stocker, M.F., Korber, G.W., Gassler, G. and Gudehus, G. (1979), "Soil
Nailing", Proceedings, International Conference on Soil Reinforcement,
Paris, Vol. 2, pp. 469-474.
Studer, J.A. and Meier, P. (1986), "Earth Reinforcement with Nonwoven
Fabrics: Problems and Computational Possibilities", Proceedings, Third
International Conference on Geotextiles, Vienna, Vol 2, pp. 361-365.
Terzaghi, K. (1943), Theoretical Soil Mechanics, John Wiley and Sons,
New York, 729 p.
Terzaghi, K. and Peck, R.B. (1967), Soil Mechanics in Engineering
Practice, John Wiley and Sons, New York, 510 p.
Vidal, H. (1966), "La Terre Armee", Annales de l'Institut Technique du
Batiment et des Travaux Publics, No. 223-234, Jul-Aug.
DESIGN OF REINFORCED SOIL RETAINING WALLS:
ANALYSIS AND COMPARISON OF EXISTING DESIGN METHODS AND PROPOSAL FOR A
NEW APPROACH

J.P. GOURC, A. RATEL, Ph. GOTTELAND

IRIGM, UNIVERSITE DE GRENOBLE, FRANCE

1. INTRODUCTION

The study presented here is a comparative study of several

methods for designing earth retaining walls with geosynthetic reinforce7
ment. In this study, only the methods considering limit equilibrium
analysis will be considered. Other methods using numerical models for
simulation (finite difference method and finite element method) are
excluded, because they are generally too laborious to design an
economical job. Moreover, it appears difficult to numerically simulate
the very variable technological conditions during construction (however,
it is one of the essential advantages of finite element method which can
simulate the construction stages such as layering and fixing the fabric
layers to the face of the wall, compacting the soil, etc.).

Since the first presence of the method "Terre Armee" (Schlosser,

et al. -(1)-), then that of bolting and geotextiles, numerous methods
for calculation at limiting state have been proposed. Without stating
all of them, they can be classified according to their hypothesis.
These consider principally the following points:

- external equilibrium
- critical slip line
- local equilibrium of an isolated reinforcement

At the IRIGM Laboratory of Grenoble University, we carried out a

comparison of these methods. In addition, IRIGM, in cooperation with
LCPC, developed a specific method, called "displacements method", for
designing geosynthetic walls.

H ( a,<p) FIGURE 1:
r.6.H Typical case

459
P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 459-506.
© 1988 by Kluwer Academic Publishers.
460

In order to simplify the problem and the comparative calcula-

tions, it is necessary to limit the number of variable parameters.
Thus, the typical case considered is shown in Fig. 1.

- The problem is considered as a plane strain problem.

- The reinforcement layers are of the same length L, regularly
spaced (bH), and horizontal.
- The reinforcement layers are composed of the same geosynthetics
(tensile modulus, J, tensile strength, TF' soil geosyn-
thetic friction angle: tan ¢g = 2/3 tan ¢).
- The soil is homogeneous for reinforced mass and the backfilling
material: cohesionless, unsaturated (y, ¢).
- The foundation soil has a sufficient bearing capacity in all
cases studied (deep failure passing under the base of the wall
will not b~ examined).
- The conditions of the wall face are not specified.

- Horizontal movement.

FIGURE 2: External equilibrium

2. EXTERNAL EQUILIBRIUM

The observations of movements of whole reinforced soil masses,

either on reduced models or on large scale models, justify the overall
study of a reinforced mass. Hence the classical procedure for analyzing
retaining walls will be used (Figure 2):

Overturning stability: factor of safety FR

• Sliding stability: factor of safety FG
Bearing stability: factor of safety Fp
Condition of no tension induced beneath the base of the wall.

This study does not bring in the characteristics of geosynthetic

reinforcement, except the interface condition at the base of the mass
(¢b = ¢ or ¢g). The calculation obtains a predesign of the wall:

L = L min (H,f3)
 461

This traditional procedure is not consistent with that for study-

ing internal stabil i ty of walls. Instead, the latter uses the concept
of slip line associated with a factor of safety Fs on the shear
strength of the soil (Fs = tan <I>/tan/<I>c) as that in limit equili-
brium analys is of slopes. While for retaining walls, the f actor of
safety focuses on the loads applied (FG,FR).

The stability of sliding can thus be examined from slip lines

such as those shown in Figure 3.

FIGURE 3: Proposal for sliding stability study

This procedure obtains continuous results between internal and

external stabilities if, for example, the same mechanism is considered
for both cases. Besides, this calculation allows a direct estimation of
the force on DC without laborious use of the tables of acting force.
However, this procedure generally slightly underestimates the active
force.

It s40uld be noted that this kinematic of failure is equivalent

to that of sliding of a wall on its base, but is not equivalent to that
of overturning. In reality, the condition of overturning is practically
satisfied at the base (FR L 1. 5) when FG L 1. 5 except for vertical
facing. Therefore, its systematic verification is not generally
necessary.

3. PRINCIPLE OF GEOSYNTHETIC REINFORCEMENT

A retaining wall resists the force P exerted by the backfill

(active block ABK) as shown in Figure 4 .
...... ---~~
6 (-P, W, R)
~
= 0

Similarly, i f the retaining wall is replaced by a reinforced

mass, the whole reinforcement will have to resist the thrust of the soil
P (active block ABK) (Figure 5).
462

FIGURE 4: In£luence of a retaining wall on the slope stability

FIGURE 5: Influence of reinforcement layers on the slope stability

---
--
~ ........ ~~
"Z (_pI, W', R') = 0
.....6 (P I,
~
~T j ) = 0
......
All the present approaches of design calculation of reinforced
soil structures are thus based on the principle of limit equilibrium:

* All the methods assume a slip line which is the position of the
maximum tensile stresses in the reinforcement (Juran et al.
-(2)-). This corresponds to the inversion of the directions of
tangential stresses along the soil-geotextile interfaces at the
points of kinematic discontinuity. Tj is the maximum tensile
force in the reinforcement j.

* Each layer of reinforcement is considered as having no flexural

rigidity (like a membrane) and only mobilizes tensile forces.
In most methods, it is assumed that the displacement velocity
along the slip line does not influence the orientation of
 463

tensile forces. Their directions are assumed to follow the

initial horizontal direction of reinforcement (~j 0, see
5.2.2) .

4. THE PRINCIPLES OF LIMIT EQUILIBRIUM METHODS

In one of several respects, the various limit equilibrium methods

are different in the shapes of the failure surfaces (slip line) assumed.

4.1 Polygonal slip line

The most general slip line is a concave polygonal slip line. The
class ical procedure of analys is cons ists of cutting the sl iding mass
into n slices. The number of the reinforcing layers cut by the slip
line is represented by r (Figure 6).

FIGURE 6: Analysis of reinforced slope stability by the method

of slices

Unknowns Eguations
Xi' Zi' e.
l
3(n+1) Xo Zo eO =
° 6
N.
l
n X
n
Z
n
e
n
°
F 1 equilibrium equations 3n
s

T. r
J
464

In the stability analysis for slopes without reinforcement, there

is a total of 3n equations obtained from equilibrium conditions of n
slices. Hence, it lacks (n-2) equations (number of unknowns - number of
equations).

In order to solve the problem, it thus needs (n-2) supplementary

assumpt ions. A method called "perturbat ions method" (Raul in et al.
-(3)-) will be used here. It consists of choosing a distribution of
normal stress along the slip line:

The above assumption supplies n supplementary equations plus two

unknowns ~l and ~2 to be determined, i.e. (n-2) equations given. The
example presented in the following text uses:

In the case of non-reinforced slopes, this method can obtain the

factor of safety of the shear strength of the soil~ F s' It can also
obtain a realistic distribution of normal forces Ni at the base of
each slice. However, it can not solve the problem with r tensile forces
Tj of reinforcements. Therefore, Tj must be determined separately.

4.2 Circular slip line

For a non-reinforced homogeneous soil mass (c, ¢, y constant), a

method of overall equilibrium can be used (without cutting the mass into
slices). A complementary assumption, eg., on the resultant force R at
the base of the slip line (Biarez -(4)-), allows Fs to be obtained
(Phan et al. -(5)-).

The forces in the reinforcements Tj need to be determined

separately, as previously, if the factor of safety for the reinforced
soil mass Fs is to be determined. The use of the method of slices in
this c~se can also be examined (cf 4.1). However, there are some points
which need to be noted in using the method of slices, say "inexact".

Bishop's simplified method -(6)- only considers vertical force

equilibrium of slices and overall moment equilibrium with respect to the
center of the circle. The shape of the sl ip 1 ine cons ide red must be
circular.

Janbu's method -(7)- can analyze both circular and polygonal slip
lines: vertical, horizontal, and moment equilibrium of each slice are
satisfied, with the exception of overall moment equilibrium (Ratel
-(8)-). It can thus be used to analyse a polygonal slip line.

In addition, there are several other different circular slip

1 ines such as a sl ip 1 ine composed of an arc and a straight 1 ine
corresponding to the sliding plane either along a geosynthetic layer
(when ¢g < ¢) or along a preferential slip line (for instance along
the wall base, when ¢b < ¢) (Figure 7). Werner et al. - (9) - proposed
 465

a particular slip line considering local instability between reinforce-

ment layers (Figure 8).

LINE

FIGURE 7: Bi-component slip line

j-J<--_Z

rupture diagram

Z approximated by string polygon

FIGURE 8: Modified slip line -(9)-

4.3 Log-spiral slip line

This method assumes a logarithmic spiral shape using the equation

below:

R R • eatan<l>c
o

(Juran et al. -(2)-, Figure 9; Baker et al. -(10)-; Leshchinsky et al.

-(11)-, Figure 10). It can obtain analytically a distribution of normal
stresses along the slip line. In the absence of reinforcement, Fs can
be obtained without assumptions. For the case of a re inforced mass,
Tj must be predetermined as in the previous cases for obtaining Fs.
466

o
ASSUMPTIONS:

• Limit state on failure

surface: t' : O"tan 4>

• Maximum plastic work

failure surface,log
spiral
• Kinematical condition :
<Xo: 1T /2

\. - - mox-~(jsin(<x-~ }d z
.,... -0 T.
AH .
Koter '. equatIon __ rr

FIGURE 9: Log-spiral slip line -(2)-

FIGURE 10: Log-spiral slip line -(11)-

4.4 Two-block

This method is very useful for analysing reinforced soil struc-

tures. The active zone is composed of two blocks. The interblock face
is either vertical (Bordairon -(2)-, Figure 11; Jewell -(13)-), as in
the· method of slices, or inclined (Hamilton -(14)-, Figure 12).

In the method of slices the number of equations required is (n-2)

(cf 4.1). Even in the two-block case (n=2), the problem still can not
be solved, i.e. the position of Na and Nb at the base of blocks
(Figure 11) must be specified.
 467

lW;

FIGURE 11: Two-block stability analysis -(12)-

D D c

_____ ~~.T,
a N,
!L--'----

FIGURE 12: Two-block stability analysis - modified cutting

of the active zone -(14)-

The "two-block method" is thus generally used without verifying

the moment equilibrium.

Unknowns Equations
2 equilibrium equations 4

Nab' Tab 2

F
s
468

The usual procedure consists of considering the case when inclination of

interblock face 5 is most unfavourable.

Therefore, in the absence of reinforcement, the problem can be

sol ved by fixing 5, then F s will be determined. The re inforcement
only intervenes by its sum ~Tj. If ~Tj is determined separately,
then Fs can be obtained. Inversely, Fs can be fixed to calculate
~Tj .

4.5 Coulomb wedge

The slip line assumed is a straight line inclined at a with

respect to the horizontal plane.

Unknowns Equations
N equilibrium equations 2
a

F
s

In this method, moment equilibrium is not satisfied either. As

previously, Fs can be determined according to ~Tj. Broms -(15)-
proposed a modified shape of active block to take into account a soil-
geosynthetic unilateral sliding (Figure 13).

ACTIVE ZONE
FOR LAYER n
ACTIVE Z9NE
FOR LAYER (n·1)

D(n·1)

FIGURE 13: Proposal of modified Coulomb wedge for reinforced

soil -(15)-
 469

5. TAKING INTO ACCOUNT THE FORCE OF REINFORCEMENT

As it has been mentioned in Chapter 4, the stability analysis of

reinforced soil mass does not allow to determine simultaneously Fs and
the forces Tj' or their sum ~Tj' In order to solve this problem,
two different approaches, global approach and local approach, will be
distinguished.

5.1 Global approach

It estimates the mobilisable strength in each reinforcement:

min

where T fj failure tensile strength of the jth geosynthe-

tic layer

FT = factor of safety with respect to tensile failure

allowable tensile strength of geosynthe-

tics

T . maximum anchorage force of the jth reinforce-

aJ ment

L .
aJ
T . 2 • f (0
n
• tan ¢g ) • dL
aJ
o

factor of safety with respect to soil-geosynthe-

tic friction

L .: anchor length (to the slip surface) L . in

aJ aJ
active zone (La.) and passive zone (L P .) will
aJ aJ
be distinguished

o : normal stress acting on the reinforcement in the

n
zone of anchorage

In order to control a design, the following procedure is

followed.

Assume that a reinforced mass of fixed length L and of predeter-

mined reinforcement, identical for all the reinforcement (Tf fixed),
the choice of FT will depend on the conditions of the use of geotex-
tiles and of its nature (Mir Arabchari -(16)-, Andrawes et al. -(17)-).
470

Once the slip line has been selected Taj can be calculated.
With Ff fixed, the mobilisable strength in each reinforcement, Tj
mob' can be obtained:

T.
J

The next step is the determination of F s' by 1 imi t equil i brium

method.

However, some variations exist. Another less physical viewpoint

such as Leschinsky -(11)- does not consider any limitation in anchorage
strength (reinforcement of large length) and assigns the same factor of
safety to both shear strength of soil and the tensile strength of the
reinforcement:

F
s
.... Tf

F
s

The limit equilibrium method associated with variationally

extremisation can obtain (Fs = FT)min.

In the most general case, Tj mob varies with the reinforcement,

and also with the anchor length Laj' and therefore with the position
of the sl ip 1 ine. The cr it ical sl ip 1 ine is the one corresponding to
(Fs)min for a given reinforcement (L, Tf/FT). Hence, the calcu-
lation of limit equilibrium must be performed for a great number of
potential slip lines.

One important remark: in order to facilitate the calculation,

many other authors follow a more rapid simplified process. It consists
of choosing independently the slip line corresponding to maximal hori-
zontal force P (Figure 4) and the length L of the reinforcement layers.

- ----
To equate Pm ax and the tensile forces in the reinforcements:

P + ~T. = 0
max J

In the case of vertical wall (f3 = 11/2), the maximal horizontal

force corresponding to the Rankine block is:

P 1 K • Y • H" where K tan" (lI

2 a a 4

This kind of slip line is most traditionally admitted ("Standard"

method - Gourc et al. -(18)-; Murray -(19)-).

For inclined walls, Jewell et al. -(13)- do the same thing: the
slip line corresponding to the force Pmax .

However, Segrest in -(20)- demonstrated analytically that the

inclination a of the critical block was not identical for non-reinforced
soil and reinforced soil.
 471

Non-reinforced soil -(Figure 14)-

Pm ax for sin 2a = sin 2(a - ¢c)

.II ¢c
-+-+ a = +
4 2

Reinforced soil -(Figure 15)-

Tf
(same TI for each reinforcement layer)

The limit value of TI to equilibrate the block for:

sin 2a - A ~H cos 2 a = sin 2(a - ¢ )

L c

¢c
-+-+ a < .II +
4 2

Note that the critical angle a depends on L, i.e., on the length

of embedment.

Therefore, it should be remembered that the critical slip line

can not be determined without taking into account the reinforcement.

/ /

'vi 1 II
-
--------~~-
-P

/,c
Q:'
/

-
R

FIGURE 14: Static equilibrium for unreinforced soil

472

FIGURE 15: Static equilibrium for reinforced soil

Moreover, even in the case where the global approach is rigorous-

ly performed, without precedent approximation, this approach is not
conservative because equilibrium is obtained from maximal mobilisable
strength (Tj mob with factors of safety FT and Ff):

~ T. ~ min
J
j j

However, nothing guaranties a simultaneous mobilisation of r

forces Tj mob as this is examined in the calculation. It's all the
more true for a relatively deformable structure.

5.2 Local approach

Most other design methods, aware of the inherent uncertainty in

the global approach and the approximations made on the critical slip
line ~the shape of the line or line of maximal tension), have completed
the precedent equilibrium condition by a local equilibrium condition for
each layer of reinforcement. Two types of approaches are distinguished:

- equilibrium of a horizontal layer of soil-reinforcement compo-

site;

- equilibrium of a reinforcement considered as an anchored

membrane.

5.2.1 Equilibrium of a composite soil-reinforcement layer

Juran et al. (Figure 9 -(2)-) showed that the vertical stress

Oz -and horizontal stress Ox are principal stresses on a central
plane between two reinforced layers because of symmetry. Consequently,
the shear stress c on these planes is zero -(21)- (Figure 16).
 473

,..- -+-- ~
T J r TENSION IN
REINFORCEMENT

~------~-----~
...... T"=O crt ..........
IT ,"7 }"7~"7

~~ ~~~______-r____U._3____________
ACTIVE RESISTANT
ZONE ZONE

FIGURE 16: Local equilibrium -(21)-

We can then consider that the active zone is composed of horizon-

tal composite layers maintained horizontally by the forces exerted along
the slip line (Figure 17).

Each composite element is subjected to a force (ox • ~H) and to

a tensile force Tj. From local equilibrium,

• ~H

This is the classical principal of local equilibrium.

Next, the various methods considering the values of Ox take

into account:

the compatibility which overall equilibrium imposes:

H
f ax . dZ p
o
,
This equation is checked in "Standard" method (method C) appl ied to
vertical walls -(18)-

<Pc
since a Ka • Y • (H-z) where Ka tan2 (lI -)
x 4 2

H
f ax . dZ 1. K • Y • H2
2 a
a

in agreem~nt with

p p 1. Ka • Y • H2 for a TI/4 + <p/2

max 2

(refer to 5.1)
474

This equation is also checked in Jewell's block method -(13)-

applied to inclined walls:

p
K • Y • (H - z) where K ~
o
X 1
-yH 2
2

p 1. K • Y • H2
max 2

,
/
"""'""--------"')--..
---=-~ -- --/---=t.--
-W ' T 'J
. . . --~7--+
---- --/~---------

-- --,--.

p -'I~R
L
- .....

1,I
-
'} I

..

FIGURE 17: Compatibility between local equilibrium and overall

equilibrium for active zone

It is clear that these r conditions of equilibrium for the r

layers of reinforcement impose safer conditions than only using overall
.
equilibrium condition ( 5.1)

T .
local equilibrium: 0xj • ~H ~ min (T I , ~) ~~ TI2
s

T .
global equilibrium: 21.)
F
s

This can compensate partially or totally the simplification on

the critical slip line identified from the line of maximal thrust seen
in 5.1. Moreover, the condition of overall equilibrium seen in 5.1 is
usually not used, the slip line goes out of the reinforced soil mass, at
the top of the wall.

Other authors obtained a result even more conservative in over-

estimating the force P obtained from overall equilibrium:
 475

-(22)- (Steward et al., - method [D]) o Ko • Y • (H - z)

X

where K 1 - sin ~c (soil at rest condition)

o

-(15)- (Broms, - method [E]) Ox 0.65 • Ka • y • H

stress distribution used in tieback walls

and K tan 2 (l! _ ~c)

a 4 2

In the case of "reinforced-earth" walls with vertical facing the

reinforced backfill is considered as monolithic: one obtains a horizon-
tal distribution of nonhomogeneous vertical stress 0z' owing to the
moment of overturning (Figure 18).

-(19)- (Murray, - method [B]) trapezoidal distribution of Oz

The thurst is increased in considering maximal stress

(oz)max

6M
o K [y(H-z) + _z_]
x a (L) 2

where K tan 2 ( l! _ ~c) and M the resulting moment.

a 4 2 z

-(21)- (LCPC - method [A]) distribution of Oz from Meyerhof

y(H-z)
o K
x a 2M
z
1 -
y(H-z)·(L)2

#(-If/
WlTIJJJJ
crn JI
(C)

H
LLIJ (8)

~
I

1//-(/1
IJIDJJ 2. ( A)

FIGURE 18: Distribution of vertical stress at geosynthetics

level
476

5.2.2 Equilibrium of geosynthetic considered as an anchored membrane

- the "displacements method" -

5.2.2.1 Basic principle

The notable difference existing between a metallic reinforcement

("Terre Armee": Reinforced Earth) and a usual geosynthetic reinforce-
ment is due to the larger deformability of geosynthetic with respect to
the soil strains. In this circumstance, mobilisation of tensile forces
in the reinforcement will depend on the displacement that it is imposed
along the slip line.

The proposed mechanism is shown in Figure 19 (Gourc et al.

-(23)-).

The displacement of the active block correspond with displacement

~e at top of the wall. A displacement field is imposed in soil in the
vicinity of the slip line, i . e., ~ej = f(~e) at the jth reinforcement.

Once the - behaviour of each reinforcement is defined for a

membrane anchored at both sides, in the passive and active zones, Tj
g(86j)' the relation 86 j = f(~e) allows a relation of compatibility
for tensile forces Tj mobilized in each reinforcement to be found.
Thus, during the displacement of the block (~e increasing), the inclina-
tion Bj and the tensile force Tj enlarge.

DISPLACEMENTS
METHOD

- " -"
sheetj
H

z
SOIL MASS

slip line

FIGURE 19: "Displacements method" -(22)-, -(25-)

 477

For each value of n9, there are r couples (Tj,Bj) obtained

from local equilibrium of each reinforcement. These r couples are then
substituted into the general equilibrium equations of the block (here,
the "perturbations method" -(3)-). The process is repeated by incre-
ments of n9 until a system of r couples (Tj,Bj) stabilizing the
sliding block is obtained.

5.2.2.2 Local behaviour of reinforcement

The experiments performed at Grenoble University (Gourc et al.

-(23)-) on a small scale model of temporary road on weak soil, rein-
forced by a fabric, brought the proposition of the mechanism of an
anchored membrane.

* Anchorage: The two sides of shear zone, the reinforceme~t is

p
zones. For a geo-

passive Zone

Critical slip line

I~j<"/2-Wjl t.Zj
aM M
u~I~~""'t~~~l..I2SJ~~J.l
Bj

Anchorage Membrane --------

Anchorage

FIGURE 20: Local behaviour of the geosynthetic layer near the slip
line: anchored membrane -(22)-
478

synthetic of linear elastic behaviour under tension (T=Joe:,

where e: represents the relative elongation) and elasto-plastic
behaviour under friction, one can determine, at the
soil-reinforcement interface, the laws of anchorage behaviour

a
u Aj = v (T aj , La. ) active zone
aJ

P
u Aj = v (T aj , LP . ) passive zone
aJ

which relate the displacement of points A~ and A~ to the ten-

J J
sile force Tj in the central zone of membrane (Figure 20).

* Membrane: This concept has been clarified from the shear test
on a soil-geosynthetic composite, the direction of shear being
perpendicular to the fabric layer (Gourc et al. -(24)-) (Figure
21).

The geosynthetic takes its shape like a membrane in the vicinity

of the shear surface, due to its negligible flexural rigidity. It
supports an increase in stress llqj normal to its plane (Figure 20).
In order to simplify the problem, llqj was taken to be uniform. This
implies a bi-circular deformed shape in the vicinity of the slip line,
at least for 119j is sufficiently small (~j < aj). For larger
displacements (with regard to soil stiffness Ks ), the membrane becomes
tangential to the sliding surface (~j = aj).

FIGURE 21: Shear test on a soil-geosynthetic composite

sample -(24)-
 479

It is considered: K llz./4
s J

B*
j
equilibrium of the membrane: T.
J 2 •

elastic behaviour of the membrane:

T.
a P
(u Aj + u Aj ) - B*'
J
.~
sini3.
--L
sina.
X.

-1 + 1 +
J B* - (LIZ j - X. )
J J
. cotl a.
J
l 0

The above equation, which integrates the displacements at the

P
head of anchorag~ (u Aj ' u Aj ), assures continuity between the two zones
of anchorage and the zone of membrane.

This method is very versatile because it can envisage an anchor-

age in nonhomogeneous soils, or particular conditions such as the
geosynthetic fixed at the wall face (uM ~ 0 for a motionless facing -
Figure 19). This method was used as an operational design program,
perfected jointly with LCPC (Cartage Program - Delmas et al. - (26)-) .
However, at the present stage, it appears difficult to determine experi-
mentally the soil stiffness Ks' Hence, the behaviour of the membrane
is simplified as follows (Figure 22):

"Small displacements" approach:

"La~ge displacements" approach:

• sin a.
J

5.2.2.3 Limit equilibrium study

In the general case, i.e., a polygonal slip line with the appli-
cation of method of slices (Perturbations), the number of equations and
unknowns are as follows:
480

SLIP LINE-
'/'
LAYER j
---- p
Uaj

Displacement Metho d
ond
CARTAGE:smoli displacement

SLIP LINE

LAYER J

Displacement Method \
and
CARTAGE: lorge displocement

FIGURE 22: "Displacements method": CARTAGE program, specific

hypothesis

Unknowns Eguations
equilibrium of slices 3n

F 1 X Z e 0 3
s 0 0 0

N. ~
n X
n
Z
n
e
n
0 3

III , 112 2 N.
~
N.
~o
(Ill + 112 . f(Si) n

Tj , flj 2r equilibrium of anchored membrane 2r

lie 1
4n+2r+7 4n+2r+7
 481

The solution of the system is obtained by fixing first the factor

of safety Fs = F, and determining the f:,S (or f:,Z), which brings the
active block into equilibrium (one f:,S for every slip line). The
critical slip line is that corresponding to maximum displacement in
equilibrium.

In demonstrative case, this procedure is applied in the case of

the critical slip line such as the case of Coulomb wedge (angle of wedge
is: (11/2 - a) (Figure 23). The displacements of equilibrium obtained
are compared for method [A] -(21)- and method [E] -(15)-. I t is
concluded that there exists in function of the angle of the block a
maximum f:,Z (or f:,S) for all cases.

Methode (E) Methode(A)

- -I
~
0;15 "--- 10;6'
~
9 I1 '---
11

8 0,8 I 1°,8
10

:1
I
I ~
~

7 0,8 9 10,8

r
"--- 0,8
'--
6 8 10.8
H-_7 m H
"--- ~
1°,8
5 7
"--- 0,8
"---
4 6 10,8
~ "---
3 0,8 5 }a,8
'--- 4
iO,4
~
10,4
2 1°,8 ~ 3
"-- '---- 2 0,4
1 1°,8 '---- 1
1
.' .•", ~
.'/,'/1
-
L-7.5m
.\ -
L-6m

~z Equilibrium (mm)

60
METHOD:

(E).---.
(A).6--6.

10 30 40

FIGURE 23: Comparison of two design methods (A -(21)-) and

(E -(15)-) by "Displacements method"
482

In the case shown in Figure 23, the design according to method

[E] obtained an equilibrium displacement larger than that according to
method [A]. Thus, a method of evaluation for one design method in rela-
tion to another is disposed.

If the "displacements method" is used as a design method, the

process is different: Delmas et al. -(26)- proposed in the program
Cartage to set a limit for maximal equilibrium displacement l1Z (or l1e).

6. INFLUENCE OF DIFFERENT HYPOTHESIS OF DESIGN CALCULATION

As it has been mentioned in Chapter 5, there are important

differences in the hypotheses of design calculation. The problem is to
know if this involves differences in design calculation. Without study-
ing all cases, we will emphasize some significant points which will let
the designing engineers know the influence of a particular hypothesis of
calculation on the design of a reinforced retaining wall.

6.1 Choice of a reference method

The simple method, "two-block" (Bordairon -(12)-, will be chosen

as a referenc.e method. It satisfies overall equilibrium (cf 5.1) but
does not satisfy local equilibrium (cf 5.2).

In general, the design calculation depends on the following para-

meters:

(Figure 24)

!
L

I /
-----~ 1---
..6 8 --I-~
-- --<:i . .----. L--
.

_, ____ ,L..__../__
[t, <DJ I '
H ---- ~-I--I----
-- - ----:,)i--.. -
~--- ~---'"' _....

FIGURE 24: "Two-block" stability: design parameters

 483

The soil is considered homogeneous, cohesionless, and tan <l>g =

2/3 tan <1>. The factors of safety for external stability are FG and
FR' and for internal stability is Fs tan <l>/tan <l>c = Ff. Once
the limiting values of FG and FR have been chosen (here FG = FR
= 1), the charts of the following type are obtained, i. e., L is the
length of the mass in function of the height Hand TI Tf/FT is
the allowable tensile strength (Figure 25).

The diagrams can be divided into three zones (with TI fixed):

The first zone is for L varies linearly with H (L = Lmin).

In this case, it is the external stability that influences the
design.

The level zone is where the choice of an allowable tensile

strength (TI) imposes a limiting height of the retaining
wall. This is the case that P = -~Tj rTI with overesti-
mated length of fabric anchorage.

The intermediate zone is where Taj/Ff < TI for some

layers.

6.2 Interdependence of different factors of safety

6.2.1 FG and F R , f(F s ) (Figure 26)

It is noted in Section 2 that it seems more logical for external

stability to maintain the same principle of study (the type of slope
stability analysis method) than for internal stability. From this point
of view, Fs 1.5 with FR = FG = 1 were chosen. Then, from the
minimal lengths obtained, i.e., LRC (for F 1, FR 1.5), LGC
(for FG = 1, (Fs = 1.5), and Lpc (for on > 0, no tensile stress at
the base of th.e wall, Fs 1.5), the required minimal external length
Lec max (LRC' LGC, Lpc) is determined. This design is then
verified for a system of factors of safety F; = 1, F~ > 1.5, F~ > 1.5 -
retaining ~alls design method). It is certain that the external length
calculated Lec is overabundant towards these coefficient F* values .
Only LRC calculated in the case of vertical walls can not obtain a
coefficient F~ > 1.5 (for F; = 1) .

6.2.2 TI and Lmin f(Fs) (Figure 27)

In Figure 27, the variation of TI and L are presented in func-

tion of the factor of safety Fs fixed for soil. Because there is no
relationship existing, a balanced choice of the factor of safety for
soil (Fs) and geosynthetic (FT = Tf/TI) remains difficult.

The minimum width Lmin and the intrinsic tensile force TI

(minimum tensile force for equilibrium) are increasing with f actor of
safety Fs.

To set, for designing, large factors of safety Fs, induce a more

conservative value of TI.
484

~=2g.2·

IPgc = (Pb
° 2 =0.
.1H=0.75 m
H(m)

/
/
/ Tr (kN/ml
20

~---17.5

~ _ _ _ _ _ 15

5 /-------125

~--------10

~-----------~

A--------------2.5

L!rr. ;

T;:lkNlml
I-fern)
17.5

/""_ _ _ ,5

f----'25
s
~------10

~--------~

~---------5

~-----------2.5

a'~~:----------:5,----~L{m)

H{m)
T;(kNlml

/ -_ _ _ 12.5

/ -_ _ _ _ _ 10

5 / -_ _ _ _ _ _ _ _ 7.5

; -_ _ _ _ _ _ _ _ 5

}-_ _ _ _ _ _ _ _ _ _ _ 25

a 5 L(m)

FIGURE 25: "Two-block method" -(12)- charts for designing

 485

'f
. .
retalnlng wa 1 1 s [ F.G =1.5
_design method FR:=1.5
Fs=1

• 1P=25°
... rp=30° ..... .....
• 1P= 40 o
. . . --.----+
"' ,
"' ,
10 ""
"'

"' "'
5 "' • ___ -e

.---~
~ -.-.-.-~-.---~-

90 60 60 90

FIGURE 26: "Two-block method": choice of the factors of safety

486

Tz (kNlm)

40

30

20

10

(]>(. )

10

20

T, (kNlm)

40

30

20

10

10

20

FIGURE 27 -a-: "Two-block method": influence of the factor

of safety FS on the TI and Lmin values
 487

TI (F,=1s)
H = 7.5 _ A ~ = 90°
TI(s: 1)
_ - . p = 60°

I
•
I
I

I
/
/
2

•
I

I
I
I

025 05 075 tg<D

. ;/ .
A

----.-- -
-_

2 A~
A

Lmin(;;=15)
Lmin(F,:1 )

FIGURE 27 -b-: "Two-block method": security on Lmin and

TI induced by FS
488

6.3 Method of blocks: influence of interblock conditions (Figure 28)

In paragraph 4.4 , it is seen that the incl inat ion Ii (Figure 24)
must be fixed first.

The most critical case (TI max) corresponds to Ii o. A

similar result has been obtained by Jewell -(13)-.

6.4 Comparison of "two-block method" with circular sliding method

The comparison is made for the following three methods: "two-

block method" with Fs 1; "perturbations method" (calculation of
corresponding Fp); and "Bishop's simplified method" (calculation of
corresponding Fb). The results are compatible (Figure 29). Note that
the "perturbations method" is an exact method which satisfies all the
equilibrium equations, however, "two-block method" and "Bishop's simpli-
fied method" are inexact.

In the case where 'Pb = <Pgc = tan- 1 (2/3 tan <Pc), the method
of circular sliding overestimates Fs. This is because the slip line
of "two-block method" can follow the weak plane of the wall base, but
not for circular slip line.

In Figure 30, the critical slip lines corresponding to each

method are presented.

I
~ ! !"lM,7X\
! 7
! 7
(2) 0 0 0 ~ =60'

30
'8 qJ.
C
(f)
'C 0 0
(f)
,b
(f)
'C
(j)
'gc
(Pc (f)
'gc

29.2 A,--c.. AL--.A. 0--0 .--.

...--.
-....
.--.
(j)
'C 21 ~ 0-0

17.3 ".,-,,:,. .......... 0-.-0

20

10

H{m)
o 5

FIGURE 28: "Two-block method": influence of Ii and <Pb

values on TI
 489

F 'P.'
c
29.2 ..
Fp Fb
0
F

1.2
21 tJ.
1.2 ..
17.3 • 0
~ IJi i
tJ.

---t---------- '-- ---~------

•
f, I o
iI
0.8 0.8
H(m) , H(m)

F F
02~ 0 ; rpb=rpc
tJ.
f
'*
1.2 1.2
i ~ .1> a
, _ _ _r!l__ ~------
•
o

0.8 0.8
H(m) H(m)
0 5 10 0 5 10

F F
C)2=0 I"b = rpgc
O~O ; rpb = rpgc
t
~
a
~
1.2 1.2 l:.

tl ~
I .- - - _.i' - -~ _______ _
•
o

0.8 0.8
H(m) H(m)
o 5 10 o 5 10

FIGURE 29: Comparison of critical slip lines: "slice method"

with circular slip line and "two-block method"
490

X (m) Z(m) I F
I Cb -3 9 I 1.08
.... _-- I Cp -3 9 I 1.07

P(kN/m) T:r(kN/m)
lSc 107 11

l Sm 107 11

"Perturbations method"
"Bishop~simplified method"
Z
~=l: "Two-block method"

FIGURE 30: Comparison of "slice method" for circular slip line

and "two-block method": deviation of factors of safety
 491

6.5 "Two-block method": choice of critical sliding surface

In 5.1, it is mentioned that an arbitrary choice of the slip line

corresponding to maximal thrust Pm (Sm)' instead of choosing the
critical slip line (Sc)' may induce an underestimation of minimal
TI'

For vertical walls, Figure 31 shows (Figure 31 -a-: H 4.5 m

and Figure 31 -b-: H = 7.5 m):

- the slip line Sm for the maximal thrust Pm ax = Pm;

- the slip line Sc for the critical thrust Pc corresponding t.o

the maximal value of the tensile force TI'

For a vertical wall, the Sm slip line (for the maximal value of
the thrust Pm) agrees with Rankine wedge. Sm is independant of the
reinforced retaining wall width L.

On the other hand, the Sc slip line (for the critical value of
the thrust Pc is a funct ion of the width L. As a matter of fact, the
computer calculation provides two slip lines for every value of the
width L (Sci, "inferior", and Scs, "superior") and for the same maximal
value of TI'

The larger the L value, the larger is the critical thrust value
until the maximal thrust Pm is reached (superabundant anchorage length
Laj). Also the larger the L value, the smaller is the TI value until
its minimum for Pc = Pm is reached.

However, it should be noted that in the method of reference (13),

the unconservative hypothesis slip line Sm corresponding to P = Pm
instead of the slip line Sc, is compensated by the added condition
related to local equilibrium.

6.6 Choice of local equilibrium conditions

For vertical walls, we compare the values of the tensile forces

TI computed by methods A,B,C,D (see 5.2) and also method G, the "two-
block method". To allow a practical comparison, the same factor of
safety is used for all the five methods.

In Figure 32 -a-, FG FR 1, Fs 1.5. The selected

width of the reinforced retaining wall is the minimum width L = Lmin,
again identical for all the five methods.

TI values are related to local equilibrium conditions and

increase with the value of the horizontal stress ax,

In Figure 32 -b-, the conventional design calculations for

retaining walls choose the following factors of safety:

F = 1 (6.2.1)
s
 492

"& ff (kN/mJ If!:. = 29.2'

IPgc = 'lib
r
f5 = 90' H I lLlH
I
62 =0'
4H=0.75 m r
H = 4.5 m L

20 --- T:{ Sc l
f--+ T,{Sml
"'---J.. Pm FIGURE 31 -a-
n
• p.

\•
".--II~I
n "Two-block method": deviation of
0---0 ~ TI values and critical slip line
15
" wit h the choice of thrust (Pm or
Pc) and the retaining wall width.

+,~
I +--- ~
H 4.5 m

A--A--- -t---.:~----·CV--:-*- ._-

10
.,11
/ .// ---' ------
/.iil ......
~.-cY-/
'" /'•
5
2 J 4 5 Um)

=
___ 5 e i
H 4.5 ____ sc:.s
_ _ Sm

L= 3 m

\ ~;=64
\ Pc,=70
\

L:3.5m 1 L= 4 m

\ Pcs=70
~;=70
\

\ \!
L:: 5 m
 493

If!: = 29.2'

I
~
'Pgc = rpb

.\
30 ~ =90'
°2=0'
~H=0.75m

= 7.5 L

\
H m

~~(Sc) FIGURE 31 -b-

25
- . . T,(Sm)
\
A---A. Pm "Two-block method": deviation of

·1'." ""'. :.: ;:,

n TI values and critical slip line

.--.----------
."", Poi with the choice of thrust (Pm or
Pc) and the retaining wall width .
• ___ _____ n H = 7.5 m

20
A __ A __ A ___ .A ____ ....~~
./
~:/
/~ /.-:?"
0.... ./ -'\iiI:;::>''-
r:/ ./
.
1S .'o....
. /" /
./ "" /
/
1/
•
v+---~--~----~--~----~--~--~-
5 6 7 Um)

\ \.
Pci.=-115 '\'. Pd~135
Pes=161
\.. Pc.=153
\
\ \' ...
\ "... "',

''':':'''
'"":~..

Lmin= 4.4 rn L:'.7m

L::6m

Pei=150
Pc.s=150 \. \
"
\ Pei=191
\PQ2194

\
\
\
\

L= 5m L::: 5.5 m L=1.5m

494

D
~D = 40
'Pg = I"b
Ti(kN/ml f3 =gO'
0,= o·
.1H:O.75m
-,,,,~,, LmfJ1

50

10

2 J 5 6 7 H(ml

A +-- r~~ ARME£

5 8 {I-.-rRRL
ABeD
C A ___ -.5TANOAFiO

o a-- __ FORCsr.sDN/CE:

IRIGM G ....... ··· .• ·-DOUBL£ BLOC G

Lmin (ml

FIGURE 32 -a-: Comparison of traditional design methods with

"two-block method"

32 (a): methods A,B,C,D

F = 1. 5
s

-G, "two-block method"

F = 1. 5
s
 495

tp = 40
'P" ='Pb i - - - - ll ,
~ =90' ~==1~ilH
0,=0'
LlH=0.75 m
Lmin

50

10

2 J 5 6 7 Him)

~~
'~
...... .....
...........
, ..., ..... ," ,
A +- - TERM: ARMEE
"
' .....
5 B {l-.-TRRL
ABeD "b= FR=1.5 F,,=1
C ll ____ STANDARD

o a----FOR£STSERV/CE

JRIGM G )...········~OOUBLE: BLOC G IS = ~= 1 r=;. =1.5

Lmin! m)

FIGURE 32 -b-: Comparison of traditional design methods with

"two-block method"

32 (b): methods A,B,C,D

F = 1. 5
s

-G, "two-block method"

F
s
= 1. 5
496

We take the example of last Figure 32 -a- in reference the "two-

block method" (for FG = FR = 1. 5 Fs = 1). rf Figure 32 -b- is
compared with Figure 32 -a-, then for methods A,B,C,D, Lmin and Tr are
smaller than those in the previous case (Figure 32 -a-).

Tr values are closer to those by method G.

A further comparison for the case of Figure 33 shows that a

deviation on ax values introduces a deviation on Tr values. The
parameters used for Figure 33 are those of the RMC Trial Wall:

H 3 m L 3 m y 17.9 kN/m
2

43° <j>
g
= 43° c = 0 c
g
o

Fs = 1 (except for "method of slices": Tr fixed)

6.7 "Displacements method": influence of complementary parameters

6.7.1 Anchorage behaviour

For the case shown in Figure 34, the influence of the parameters
up, La' and J on Ta f(uA) is presented three-dimensionally in
Figures 35-37.

(Figure 35) up has little influence.

(Figure 36) La influences little away from the maximal

anchorage strength.

(Figure 37) On the other hand, J has a notable influence.

6.7.2 Membrane behaviour

, The variation of soil stiffness Ks induces a modification of

equilibrium conditions.

While Ks increases, the equilibrium displacem~nt 6Zj

decreases and ~j (the inclination of tensile force Tj) ~ncreases
(see Figure 38).

Note also that simplification of the geometry of membrane zone

(Cartage program - see 5.2.2.2) in tensile force (Figure 39: case of
Figure 33 - Q = 50 kPa, "small displacements" approach).
 497

0=12 kPa
10 20 30 Ti 20 40 60 ax
(kN) (kPa)

II: \1\
~I:
\\: I

\\t~ \\\ \ \\ \\
\\ \

t\ \ \1 ~
c~rcle( Fs:::;1.03)
x=_6.75
z"" 4.25
\l ~\

0= SOkPa

10 20 30 40 Ti ,I
1/
(kN)

i
W
\\ 1\ \\
l\ ~
IIi doubl~ bloc I
It /[ t

I
T\
~\ \'\ 1\ \\
1\ '. .Ii)
1: coin Ronkin. \

:! 1 J

\\ \\ ILl cercle (Fs= 1.04) I

I
J

ii \
I
I \ \ ~ I' x=-8.7S
z .. 5.25
!
V

0= 1QOkPa
A +- TcAAC ARI-IEE

B (t---TRRL
C 6-~---STANDARO

0 . ____ FOREsr.sERVlr.E.

G J..·········-OOUBLE BLOC IL:::===L_L

10 20 30 40 Ti 20 40 60
(kN) (Jd'b)

•• fL,/ i
1/
~r
L \\ \
\
\
,,"
double- bloc
I,
Iii \1 \
I
\\
\ \I i/
\

~\. \
\
\
I-
"
\\ \ \ 1i
c~rcle ( Fs_1.04 )
~;-m
\

I
,:1
/ \\ Ii
I
\ I , COin Rankin~
,
)2'
i
I
I

I: I I \\ ,

X
i
FIGURE 33: Comparison of traditional design methods with "two-block
method" and "method of slices" associated with circular slip line
(RMC-Kingston geogrid reinforced earth wall)
 498

J
-f-.L-~---_ E!%l
Tensiletest

FIGURE 34: Anchorage behaviour parameters for a geosynthetic

buried in soil mass

200

200
J 500 KN/M 5 M 180/~'

P ~
20,78 KPA C
160,-,
7'
UO S:-
Valeur deTA

..--
~
/20
<lJ
0 HBGIJE 189 100 (J1
o
0 170 - 189 ~

LSJ 151 170

132 - 151

-
113 - 132
94 - 113
76 - 94
Ifill 57 - 76
~ 33 - 57
19 38
BELOIJ Ig

FIGURE 35: Anchorage behaviour of a geosynthetic buried in

soil mass: influence of Up
 499

.:;~\\\\\\I
-, ,39;'
(- +----------- ""
___ _ .
<~~ ! '4 2 -1 - - - -

J = 500 KrUM ,
Up = 0, 01 11
QJ 2 4 -2 ,-
0p 20,78 KPA
en
TA
2 192
V a le u r d e

-.=.
r- 14 2 242 Q)
o CJJ
8o 1l6UIJ[ JS i'
.J,:) - .1.')c
31</ - ].1;;
«
1-
92 19 2
U
oL

::
i'SO - J I</ 14 2 C
i" 16 - ,'.') f) o

-
212 - 2'10 «
17,1 - 212
92
I- -
1'1; - 17,)
~ Iff ) - 1'1,1
7,; - 110
B[LOI) 76

ri e d in
e o sy n th e ti c b u
~: e b eh av io
u r o f a go f L a
A n ch o railg m as s: in fl u e n c e
sO

Up = 0,01 M 5 M
0p = 20.78 KPA

V a le u r d e T A

CJ

:-
n6(j{J[ 19 1

=
0 17'1 191
0 /,"),')
I ,If)
// .-,
157
1('3 I'll )

:
11)6 Ie,
09 1f)6

=
72 39
~ $') . - ,(
,8
T')

.f;!;
6[LOr,) 38

th e ti c b
u ri e d in
e o sy n
~: ho ra ge b eh av io
in
u r o fn cae go f J
fl u e
A nC o il m as s:
s
500

I INFLUENCE of SOIL STIFFNESS Ks I

.1Zj

FIGURE 38: Membrane behaviour: influence of soil stiffness Ks

RAlDEUR du 50L, Ks..22600kN.M>

LOI d~ DEFORMATION:
T~ Tp='fZ.kJ'Pg+C g
J>-- --J> DOUBLE BLOC
Tp •. __ , U =00002S.~
METHODE
____ on DE
'.lOOH PlACEMENT
J=43SkN/M 4 f---J---;!!--~ : p, rz·S4100
: U(m)
~z (m)= 0,0112 J f-.,..!ll!.ll!----!~--l Up
METHODE CARTAGE L.C.?'C
o--o.=lOOH J:-43SkN/M
~z (m)= 0,003 4
Q= 50kPA
x
10 20 30 40 50 60 Ti(kN/M) 1! !
I
I

,.' DOUBLE BLOC

\ Ii
/ METHODE ~n DEPLACEM!
METHODE CARTAGE
I (x.S,OO, Z.:;OOI

, /

FIGURE 39: Comparison between "displacements method" and "two-b1ock

method": case of Figure 33
(RMC-Kingston geogrid reinforced earth wall)
 501

The following is a design example by the program Cartage, with

different assumptions shown in the table (Figure 40) (Delmas et al.
-(26)-).

10 20 IKN/'llJ o 10 IkN/ml '( I " 18kN/m 1

\ TENR TENP c. = OkPo

\

19 "Pg / Ig 'i,. :; 0,8

2 up " O.D1Sm

q " 2SkPa

,,
.. I
+ 5

I
I
I
I

101
J.. 1 Lit no) lei

TlkN/ml
20
l Cas nO 1 I2 J ,
!JlkN/ml 500 I 2000 500 500

I"deplacemenr' petit I petit petit grand

I fixe
.
I"paremenr' libre fixe fixe
10
1+ • .
Parement

-----
o Iml

Ibl

FIGURE 40: Design of CARTAGE program -(26)-

 502

6.8 Seismic effect

The "two-block method" is used again and the seismic load is

simulated by a psudo-static load (with a seismic coefficient k). The
internal seismic action is taken into account in the form of a horizon-
tal force acting towards the slope, koW. Similarly, an external seismic
action is assumed existing on the active block behind the reinforced
wall. This thrust is a uniformly distributed load acting on DC (Figure
41) (Monobe et al. -(27)-). This consideration is conservative because
both maximal values of the two effects (internal and external) are
assumed acting simultaneously (Bastick et al. -(28)-).

c K
/

FIGURE 41: Internal and external seismic effect on a reinforced

retaining wall

L(m)
IP.: =29.2·
17.5
IP.: =37.4"
S

fP9C = fPb
,

°=
, ~ =90·
I
I
I 2 =0.
5 I .1H 0.75 m
I
I kaO;1/1: (FootS)
I
I k.OjCP;

".-.w

H ~.1 H
'" L
H(m)
5

FIGURE 42: "Two-block method": seismic effect

 503

If our results are compared to those obtained from Bonaparte et

al. -(29)-, it is noted that for the same factor of safety sF 1.1,
s
we obtain the values of the width L distinctly high (the increase of the
external thrust not being taken into account in reference (29)).

By comparing with the results for static case (Fs = 1.5) , i t is

seen in Figure 42 that the protection brought by a static design is
weak in relation to an earthquake (F ... < O. i) .
s =
1.5 F S = 1.1 for k
s

6.9 Cohesion of soil of reinforced mass (Figure 43)

In the design of a reinforced soil mass, the weak cohesion of the

material is usually neglected (cc = 0 instead of Cc = C/F s ) ' This
induces a distinct overestimation of the minimal value of TI (Figure
43) .

H ~ 7.5 m
<Pc =21'

4-
I 7
iPgc = rpb H ! 7
O2 =0' / 7 L'.H
~g_--t7
L'.H=0.75 m """,.,,,[}I-' 7
Lmin

3 -- f3 =90'
..0.---.0. f3 =60'
2

o 5 10

FIGURE 43: "T' ,-block method": influence of the soil cohesion

504

7. CONCLUSION

The numerical and graphic methods will provide the engineers with
explicit charts for rapid predesign. However, it is also necessary to
know well the fundamental assumptions in each method to estimate the
design scattering.

Nonetheless, the design methods still need to be improved for

taking into account such effects as, eg., a point load acting on the top
of the wall, the folding of the fabric layers to make the wall facing,
etc. These remain to be developed. However, the recent methods have
indeed made great progress in design calculation.

ACKNOWLEDGEMENT

The authors sincerely thank Dr. Rong-Her CHEN, Associate

Professor of National Taiwan University in Taipei, for translating this
text during his stay in our research group of IRIGM, Grenoble
University, FRANCE.

REFERENCES

1. Schlosser, F., Vidal, H. La Terre Armee. Bulletin de liaison

des Laboratoires des Ponts et Chaussees no 41. November
1969.

2. Juran, I., Schlosser, F. Theoretical Analysis of Failure in

Reinforced Earth Structures. Convention ASCE. Pittsburg
1978.

3. Raulin, P., Rouques, G., Toubol, A. Calcul de la stabilite des

pentes en rupture non circulaire. Rapport de recherche des
Laboratoires des Ponts et Chaussees no. 36. 1974.

4. Biarez, J. Equilibre limite des talus et barrages en terre.

Annales de l' ITBTP. Sols et Fondations no. 51. 1965.

5. Phan, T.L., Segrest in, P., Schlosser, F., Long, N.T. Etude de
la stabilite interne des ouvrages en terre armee par deux
methodes de cercle de rupture. COllOje International sur Ie
Renforcement des Sols. Paris, 1979. ,

6. Bishop. The Use of the Slip Circle in the Stability Analysis of

Slopes. Geotechnique, Vol. V.

7. Jambu, N. Application of Composite Slip Surfaces for Stability

Analysis. European Conference on Stability of Earth Slopes,
Discussion, Vol. III. Stockholm, 1954.

8. Ratel, A. La "methode en deplacements" appliquee aux massifs en

sol renforce par geosynthetiques. These D.I., IRIGM, Univer-
site de Grenoble I, 1987.
 505

9. Werner, G., Resl, S. Stability Mechanisms in Geotextile Rein-

forced Earth Structures. 3rd International Conference on
Geotextiles. Vienna, 1986.

10. Baker, R., Garber, M. Theoretical Analysis of the Stability of

Slopes. Geotechnique 28 no. 4. 1978.

11. Leshchinsky, D., Reinschmidt, A. Stability of Membrane Rein-

forced Slopes. ASCE Journal of Geotechnical Engineering, Vol.
III no 11. November 1985.

12. Bordairon, M. Dimensionnement des massifs en sol renforce par

geosynthetiques. These 3e C., IRIGM, Universite de Grenoble
I, November 1986.

13. Jewell, K.A., Paine, N., Wood, R.I. Design Methods for Steep
Reinforced Embankments. Symposium on Polymer Grid Reinforce-
ment in Civil Engineering. 1984.

14. Hamilton, M. Calculation Method for the Stability of Reinforced

Embankments. Delft, 1984.

15. Broms, B.B. Polyester Fabric as Reinforcement in Soil. 1st

International Conference Soils, Textiles. Paris 1977.

16. Mir Arabchari, N. Fluage des materiaux textiles utilises dans

les ouvrages de genie civil. These D.I., Ecole Centrale.
Paris, 1985.

17. Andrawes, K.Z., McGown, A., Murray, R.T. The Load-Strain-Time-

Temperature Behaviour of Geotexti1es and Geogrids. 3rd Inter-
national Conference on Geotextiles and Geomembranes. Vienna
1986.

18. Gourc, J.P., Bordairon, M. Rembalis renforces par geotextiles.

Comparaison des methodes de calcul. Journee sur Ie Renforce-
menD des sols par geotextiles. Rapport interne Comite
Franyais des geotextiles et Geomembranes. 1984.

19. Murray, R.T. Design of Reinforced Earth Walls. TRRL Int.

Report. 1978.

20. Segrestin, P. Calcul d'un massif en terre armee par les coins
de rupture. International Conference on Reinforced Earth.
Paris, 1979.

21. Ministere des Transports Franyais, DGTI. Les ouvrages en Terre

Armee Recommandations et regles de l'art. LCPC, SETRA. 1979.

22. Steward, J.E., Williamson, R., Mohney, J. Guidelines for Use of

Fabrics in Construction and Maintenance of Low Volume Roads.
Chapter 5. U.S. Forest Service. Portland, Oregon, June 1977.
506

23. Gourc, J.P., Ratel, A., Delmas, Ph. Design of Fabric Retaining
Walls: The "Displacements Method". 3rd International Confer-
ence on Geotextiles and Geomembranes .. Vienna, 1986.

24. Gourc, J.P., Matichard, Y., Perrier, H. Delmas, Ph. Capacite

portante d'un bicouche, sable sur sol mou, renforce par
geotextile. 2nd International Conference on Geotextiles.
Las Vegas, 1982.

25. Gourc, J.P., Mommessin, M., Monnet, J. Geotextile Reinforced

Embankment Over Weak Soil: Different Theoretical Approaches.
3rd International Conference on Geotextiles and Geomembranes.
Vienna, 1986.

26. Delmas, fh., Berche, J.C., Gourc, J.P. Le dimensionnement des

ouvrages renforces par geotextile: Programme Cartage.
Bulletin de liaison des Laboratoires des Ponts et Chaussees,
142. Mars, avril 1986.

27. Monobe, N., Matsuo, H. On the Determination of Earth Pressure

Duririg Earthquakes. Work Engineering Congress, Vol. 9, no
388. Tokyo.

28. Bastick, M., Schlosser, F. Comportement et dimensionnement

dynamiques des ouvrages en Terre Armee.

29. Bonaparte, R., Schmertman, G.R., William, N.D. Seismic Design of

Slopes Reinforced with Geogrids and Geotextiles. 3rd Inter-
national Conference on Geotextiles and Geomembranes. Vienna,
1986.
FACTOR OF SAFETY CONSIDERATIONS IN REINFORCED SOIL STRUCTURES

R.T. MURRAY, Transport and Road Research Laboratory, England.

ABSTRACT

A comparison is given of the conventional method of design of

reinforced soil structures , involving the use of overall or ' lump'
factors of safety, with an alternative method which allows a more
realistic assessment of the contribution made by the soil. The results
of the comparisons for both internal and external stability are
presented with respect to the specified factor of safety determined in
the usual way. The study has shown that relative to the specified
factors of safety, the values calculated by the proposed method are
generally smaller, apart from at the limiting condition when they are in
agreement, provided that the assumed movement of the wall is outward
rotation about the base. For the case of outward rotation about the
top, even at the limiting condition the results generally disagree. The
study has also shown that, regardless of the fact that only one safety
factor may be specified, there are at least two different safety factors
influencing behaviour. This result thus provides support for the use of
partial factors in design as has been recently proposed.
The importance of considering strain behaviour in design is
discussed in the paper. Although the method referred to above does not
truly consider strain, it provides much greater consistency and
compatibility of soil-reinforcement design while at the same time
preserving the simplicity of existing methods. The method seems best
suited to design involving "stiff" reinforcements.
The increasing use of polymeric reinforcements has highlighted
the need for an alternative approach which considers strain behaviour
directly. One such method has been described in the paper and the
equations for internal local stability are presented. The analysis has
been limited to the simple case of linear variation of strain of the
soil and reinforcement but could be readily extended to more complex
situations.

1. INTRODUCTION

The subject of factor of safety is one which generates

considerable controversy in civil engineering (Burland et
al, 1981; Symons, 1983). In this paper the conventional approach to the

507

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 507-540.
Crown © 1988.
508

design of reinforced soil structures employed in the United Kingdom

(Dept. of Transport,1978) is compared with an alternative deterministic
procedure which is considered to provide a more· realistic basis for
design. To allow the conventional and proposed solutions to be directly
comparable, similar forms of design equations are employed in the two
cases but this is not a requirement of the proposed method. An important
feature of the method is that it permits the influence of the
constraints imposed by the reinforcement on the soil properties to be
taken into account.
The calculations involve the usual assumption in design that the
wall movements will involve outward rotation about the base. However, as
recent research has indicated that reinforced soil retaining structures
constructed with part-height panels tend to rotate about the top of the
structure (McGown et aI, 1987) further analysis of internal stability
behaviour has been also carried out for this case.
As described in the paper, the behaviour of a reinforced soil
structure is greatly influenced by the strains developed in the soil and
reinforcement. This is of particular relevance to soil structures with
polymeric reinforcement although at present the design of reinforced
soil structures employing geotextiles or related materials as
reinforcement generally follows the procedures laid down for metallic
and other types of high modulus reinforcement. However, polymeric
reinforcements are often more strain susceptible than metallic types and
to avoid excessive strains being induced it is usual practice to
increase the factor of safety such that the polymeric reinforcement has
a working load which is a much smaller proportion of its ultimate (long
term) load capacity than its metallic counterpart.
As far as is known this approach generally offers a viable
solution to coping with the greater strain susceptibility of geotextiles
and related materials but their increasing use for reinforced soil
applications warrants the development of a more appropriate method of
design which takes due consideration of the material properties. The
paper, therefore, gives consideration to the development of a design
procedure which is based on an assessment of strain in the soil and
reinforcement and may thus have better applicability to structures
employing geotextiles or related materials.

2.0 ASSESSMENT OF SAFETY FACTORS BASED ON OUTWARD WALL ROTATION

ABOUT THE BASE.

In the design of retaining structures it is usual practice to

assume that the limiting condition will correspond to the active state
produced by outward rotation about the base. The analysis given in this
Section is also based on this assumption while an alternative, and often
more realistic approach is considered in the following Section.

2.1 Adherence Resistance.

The conventional procedure for determining the required length

of reinforcement to prevent adherence failure is based on a limit
equilibrium approach. The method of assessing local stability will be
considered in relation to self-weight of the fill only but the
 509

techniques discussed are equally applicable to other more complex

loading conditions.
At limiting equilibrium the force (D) applied to a single element
is given by

D ~ Ka'YzSVSH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)

The resisting force (R) is given by :

R~2BJLL'YzLL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)

At limiting equilibrium these two equations are equal

i.e R~D

Assuming that the objective of the design is to establish a safe

length of reinforcement having previously selected values of B,SV and SH
for known values of JLL and Ka , then the above equation can be
re-arranged to give the limiting length required :

KaSVSH
i.e. LL~------- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
2JLLB

Now the length to be employed in the final design must have an adequate
factor of safety against adherence failure and the required design
length is therefore given by :

It is apparent that for values of F other than unity, there is no

longer an equality with the right hand side of Equation 3. As SV,SH and
2B are unchanged, the effect of increasing the length is to alter the
properties Ka and JLL, previously defined as limiting values, and which
is no longer the case. The manner in which these two values are altered
with increasing specified factor of safety will be governed by their
relations with soil strain. From a theoretical standpoint the extreme
range of the possible values that can occur is given by :

JLL
JL (minimum)~ <--- JL,K~ function(F) ---> K(maximum)~ Ka.F
F

This relation states that in one extreme only the interface friction is
influenced by the safety factor while the other extreme is the full
effect of safety factor being taken only by the soil friction. Between
these extremes both JL and K are affected to some extent.
510

For purposes of the following analysis it is assumed that as both

values are a function of the soil friction, the interface friction
coefficient ~L can be related to the peak friction angle by a
coefficient a:

i.e ~L~tan0~ ~ oo.tan0p

Moreover, the mobilised friction angle (tan0M) may be related to

the peak friction angle by the well-known equation employed in
slope stability analysis :

ootan0p
i.e ~M ootan0M~ - - - - - - - - .......................... (4)
FS

Thus the design length can be expressed by the relation

KMSVSH
LD~- - - - - - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
2~MB

Now Ktrtan2(45 - 0M/2) ............................. (6)

Expanding Equation 6 and substituting Equation 4 gives the mobilised

lateral earth pressure coefficient in terms of the safety factor and
peak friction angle :

J(FSZ + tan Z0p)-tan0p

i.e Km~ - ---------------------- . . . . . . . . . . . . . . . . . . . . . (7)
j(Fs2 + tan 20p)+tan0p

now as F

F ----------- . . . . . . . . . . . . . . . . . . . . . . . . . . (8)

F.Ka
and FS~ ..................•.................... (9a)

or in suitable iterative form

 511

F.K a [J(FS 2 i + tan 2 0p) + tan0p]

FS,i+l ----- -- -~------------------- ................. (10)
[!(FS 2 ,i + tan 20p) - tan0p]

thus enabling the factor of safety of the soil (FS) to be determined as

defined by Equation 4. Moreover, as the strength properties of the soil
are not fully mobilised, the factor of safety against adherence or bond
resistance (FA) does not correspond to the specified value F.

RM 2BLDJ.tM
i. e. FA
DM SVSHKM

2BLL Ka
and where LL LD/F
SVSH J.tL

............................ (11)

It should be noted that the numerical value of FA in Equation 11

is unity (apart from the effect of rounding errors) for all values of
friction angle and specified factor of safety. This occurs because with
increasing specified factor of safety the ratio J.tM/KM reduces at the
same rate to exactly compensate and preserve the equilibrium of the
equation. However, this result should not be interpreted as indicating
that the structure is at the point of failure as any tendency for bond
failure to ' occur would induce soil strain and an increasing value of
mobilised friction angle. Thus to attain the limiting values of J.t and K,
the pUll-out force required would be equal to FA times the working load
which would correspond to the ultimate force based on the specified
factor of safety F. It could well be that the strains to achieve such a
condition would be unacceptably large although this is most likely to
arise with soils in a loose condition, or with highly extensible
geotextile reinforcement.
In essence the analysis makes the assumption that with a factor
of safety of unity, sufficient yielding of the wall takes place to fully
mobilise active earth pressure conditions and interface friction. As the
factor of safety is increased the amount of yielding reduces
proportionately with a corresponding reduction in mobilised friction
angle. Thus greater force is developed which must be resisted by the
element. This behaviour is analogous to a rigid retaining wall or rigid
tunnel lining which inevitably needs to support greater forces because
of the constraints they apply to the soil. Neglecting situations where
512

passive pressures might develop, it would seem reasonable to assume that

a lower bound friction angle consistent with zero strain and
corresponding to KO conditions would apply.
Figure 1 shows the relation between the specified safety factor
and t~e value of FS derived from Equation 9a for a peak friction angle
of 30 . Similar forms of relation are produced for other values of peak
friction angle. The figure also shows the relation between specified
factor of safety and the factor of safety in adherence based on Equation
11. In the analysis the mobilised friction angle was not permitted to
fall below the zero strain or KO value. It was previously stated that
the factor of safety against adherence failure will remain sensibly
constant at about unity but an anomalous situation occurs when the
specified safety factor is further increased after the zero strain
condition has been attained.
It is important to consider how this factor of safety in
adherence is developed in order to preserve the equality between the
restoring force and disturbing force. Because of the assumption that
both soil friction and interface friction are modified identically by
soil strain, the implication of the analysis is that any. additional
length of reinforcement, corresponding to an enhancement of the
specified safety factor beyond the zero strain condition, does not
contribute to the resisting force but nonetheless provides additional
strength reserve in adherence. The increase in adherence safety factor
shown in Figure 1, after the safety factor of the soil attains a
constant value, arises from this length in reserve ..
An alternative method of analysis which has been carried out is
based on the assumption that the interface friction angle continues to
reduce with increasing length of reinforcement so that the same
resistance will be developed over a longer length. This analysis
therefore requires that the relations between soil strain and both soil
friction and interface friction, which were assumed to be the same
initially, diverge after the zero strain condition is reached. It would
be possible to employ different soil strain relations for these two
modes of friction but this would add further complexity to the analysis.
It should be noted that, although implied otherwise by the theory, even
with this latter method of analysis a stage must be reached with large
value~ of LD when further increases will not contribute to resistance.
The extra length would thus be redundant but could still be considered
an additional safety factor in reserve as for the former analysis. The
factor of safety of the soil on the basis of this latter analysis is
given by:

F.Ka
for f o FS ......................... .... ... (9b)
KO

Figure 1 also shows the alternative relation between safety factor of

the soil based on Equation 9b and the specified safety factor. The two
relations showing the factor of safety of the soil in Figure 1 may be
contrasted with the normal assumption that the soil strength will be
fully mobilised at all stages which corresponds to a constant safety
factor of unity.
The above considerations of safety factor are not necessarily
indicative of unsafe design but highlight inconsistencies in the present
 513

4 .00

Local stability assessment:

rotation about toe
3.20

.... ....
.... ....
Fs (Eqn 9a)
---~ ..... ---
........ "

...... .......;:~~O.(\ ",

"
0 .80

0.00 L _ _L _ _L _ _-'-_ _....L_ _--1._ _- J_____L.-_---'

1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
Specified factor of safety

Fig. 1 Relations between specified and calculated factors of safety

1.00

0 .80 "- ..
.. --
-. - - - - __ -. q,p == 30°
tpp,--
U
'"
.§-
!t 0.60
~
:J
- _
tpp ~ 50
--
0
~400

a.
.s
u..
0.40
'0
a
.;:;
'"
a:
0.20 Local stability assessment:
rotation about toe

0 .00
1.00 1.08 1.16 1.24 1.32 1.40 1.48 1.56 1.64
Safety factor of the soil

Fig. 2 Relations between safety factor of soil and F(rupture)/F(spec)

514

approach based on the use of overall or lump factors of safety.

Nonetheless, there may well be situations where difficulties with local
stability can arise. For example, where there is poor soil-
reinforcement interaction as a result of inadequate compaction,
overlarge fill material or other cause, the larger forces may induce
excessive deformation. However, the analyses has more serious
implications in regard to the design of local stability against tensile
rupture as described below.
An important result from the above analysis is that although only
a single factor of safety was specified, it is apparent that two factors
of safety, FS and F actually determine behaviour. This result seems to
vindicate the use of the partial factor of safety approach proposed by
Murray and McGown (1987).

2.2 Local tensile stability.

The approach adopted in the United Kingdom for assessing the

tensile strength requirements of reinforcing elements (Dept. of
Transport,1979) is based on a permissable stress method in which a
permitted value of stress for a particular material (PM) multiplied by
the minimum area of cross-section (a) is equated to the horizontal force
to be supported by the element. It is assumed for purposes of the design
calculations that the shear strength of the soil is fully mobilised.
Although the design calculations for bond resistance and
tensile resistance are treated independently, it is apparent that the
same disturbing force applies in the two cases and there is therefore a
direct interrelation between the two possible modes of failure. Thus the
imposition of a factor of safety to prevent bond failure constrains the
soil and increases the forces which also have to be resisted in tension.
However, whereas bond failure would be preceded by soil strain and hence
an increase in mobilised friction, with rupture this is not necessarily
the case, particularly for high modulus reinforcement which would strain
very little prior to failure. It is thus important to ensure that the
factor of safety against bond failure is always smaller than against
tensil-e rupture if sudden or very rapid modes of collapse are to be
avoided, especially where the possibility of corrosion or degradation
exists.
In the following determination of the safety factor against
tensile rupture, the mobilised friction angles for the soil are assumed
to correspond to those values determined previously in assessing bond
stabili ty. In some circumstances this may be considered an optimistic
assumption as the relation between force and strain will be different
for the two cases and with metallic and other forms of high modulus
reinforcement, the strains induced as a consequence of tensile force are
likely to be very small. The procedure is as follows :

The tensile resistance at the working condition (RM) is given by

where the specified factor of safety F -------- . . . . . . . . . . . (12)

 515

defining the actual safety against rupture (FR) as

a·PL
FR -------- ........................... (13)
KM'YzSVSH

Thus the ratio of specified to actual safety factor against

rupture is given by :

Now F is specified and KM is given by Equations 6 or 7. As FS is ~ssumed

to correspond to the same relation with KM as previously determined for
local adherence stability, the actual factor of safety against rupture
can be determined in terms of FS.

i.e. FR

Or expressing FR/F as a ratio and replacing KM by the relation given in

Equation 7 then:

------------------------------- ............ (14)

F

The ratio of actual factor of safety to specified factor of

safety against rupture is shown plotted against safety factor of the
soil in Figure 2 for peak friction angles of 30·,40· and 50·. It should
be noted that although the local safety factors in adherence given by
Equations 9a and 9b are different, this does not affect the results
produced by Equation 14 as KM has the same values over the entire range
for the two ' methods. The relation shown in Figure 2 thus applies to both
methods of analysis.
A point to note is that the results have been normalised with respect to
the limiting condition. Thus as the specified factor of safety
increases, the gain in FS or FR may appear smaller for soils with a
larger angle of friction.
In contrast to the previous analysis involving adherence
behaviour, with high modulus reinforcement there will not necessarily be
any further increase in mobilised shear strength of the soil up to
failure of the reinforcement.

2.3 Overall adherence resistance.

A similar approach to that described above for local stability

may be employed for the assessment of overall bond resistance. Referring
to Figure 3 then the bond resistance of the ith element assuming a plane
surface as shown is given by :
516

.. oL
'I .'

I
I

T
Sv

1
z ...L

ith reinforcement
.. ,

Nth reinforcement
,....,,0------ L -------1
.. 1

Fig. 3 Geometrical considerations in assessment of overall stability

1.00 ,..--r--------------------------...,
Overall stability assessment:
rotation about toe
0.80

<>
'"
~
!S .
0.60

'"
>
..£
u.
'0 0.40
.g
<0
a:

0.20

0.00
0.80 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
Factor of safety of the soil

Fig. 4 Relations between F(soil) and ratio of F(over)/F(spec)

 517

Ti = 2BuL)'z (SL + ztan(3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (15)

Now oL = L - H.tanf3 ; Z = i.SV

Ti = 2B~L~(iSvoL + i 2 Sv 2 tan(3) ........................ (16)

Summating Equation 16 from 1 to N can be shown to produce the

following equation :

i=N
TT = 2: T .... (17)
i=l

2 (L - Htan(3) 2 1
i.e. TT ~~B~LH .2(N + 1)[------[(--- + tanf3( '3 + 6N)] ... (18)

Now the disturbing force D = ~~H2, where K takes the value of Ka at the
limiting condition or KM according to the mobilised friction angle. The
specified factor of safety (F) is therefore given by :

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (19)

and actual safety factor (FOVER) by :

2B~ij
FOVER = -KM (N
L
+ 1) [il - -'3
JKij
] ......................... (20)

Thus the ratio of FOVER/F is given by :

~O~Ea -Fi~~~1~i~-~~'~!}3f ........................ (21)

The ratio of FOVER/F has been determined from the values of KM and FS
calculated in the local stability analyses employing Equations 7 and 10
respectively, The results are presented for peak friction angles of
30°, 40° and 50° in Figure 4 for geometries of L/H of 0,6 and 1.6. As
can be seen the results are very similar for these two cases in view of
the fact that the data have been normalised with respect to the limiting
conditions.
It may be noted that the bracketed terms in Equation 21 have very
similar numerical values and a close approximation to the safety factor
against bond failure is given by :

F,K a
FOVER
518

The above equation corresponds to that derived for local bond stability
(Equation 11) and the agreement between the two possible modes of
instability can be explained by the fact that the same relation between
specified factor of safety and mobilised friction angle has been assumed
for both cases. In practice, the mobilised friction angle may vary
throughout the height of the wall according to the magnitude and form of
yielding that takes place. However, the analysis which has been
described above relates to the most commonly assumed case of outward
rotation about the base.

3.0 SAFETY FACTORS FOR INTERNAL STABILITY ASSOCIATED WITH WALL

ROTATION ABOUT THE TOP.

There are essentially two forms of reinforced soil walls; the

most common form is constructed with part height panels which permit
considerable articulation during construction. The other type involves
the use of full height panels which are generally propped during
construction. Because of the different construction techniques used for
the two methods, the strains developed in the soil are best represented
by rotation about the top in the former case and by rotation about the
base in the l.atter (McGown et al, 1987). These differences in the
pattern of strain distribution have a significant influence on the
magnitude and distribution of the lateral pressures.
A method of assessing the influence of different forms of wall
movement on the lateral pressures developed behind retaining walls has
been proposed by Dubrova (Harr ,1966). For the usual assumption of
rotation about" the base, the method produces the conventional Rankine
solution. To evaluate the lateral pressures produced by rotation about
the top, it is assumed that the mobilised friction angle increases
linearly from a minimum value at the top to the fully mobilised value at
the base. In the analysis presented by Dubrova, the value at the top was
assumed to be zero but it seems more appropriate to use the value
corresponding to "at-rest" conditions. On this basis the variation in
the maximum friction angle which can be mobilised (0PZ) is given by

0pZ = 00 + '1z . . • . . . . . . . . . • • . • . . • . . . . . . . . . . . . . . . • . . . . • (22)

where '1 = (0p - 00)/H

A further assumption involved in the method is that the total force (P)
at the limiting condition conforms to that produced by the usual Coulomb
analysis. The lateral earth pressure is determined by differentiation of
the total force equation and treating Kaz as a variable :

i. e. P = ~Kaz)'z2

dP d
i.e.
dz Kaz)'z + ~)'z2·dz(Kaz)

now
 519

i. e.

For the case of wall translation, Dubrova proposed that the

lateral pressure distribution was obtained from the average of the
distributions determined for rotation about the base and the top.
The above method of analysis has been employed in the assessment
of internal stability involving rotation about the top.

3.1 Local stability assessment.

At the overall limiting condition the mobilised friction angle is

assumed to be given by Equation 22. It is apparent that in the ul?per
part of the reinforced soil wall the friction angle will not attain its
peak value even at the limiting condition. This may be contrasted with
the case of rotation about the base when the strength of the soil was
fully mobilised at the limiting condition. It should be noted, however,
that if the outward movements of the facing continue beyond that
required to achieve the limiting condition, further redistribution of
lateral pressure would occur and eventually attain the full strength of
the soil at all locations.
As for the case of rotation about the base, the influence of the
safety factor will be to constrain the soil. Proceeding with the
analysis as previously, an equivalent equation for local bond stability
for rotation about the top is as follows :

F - - - - - - - - - - - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (24)

Where the subscript m indicates mobilised friction while the subscript z

indicates variation with depth.

Defining 00 ~ 00(1 - z/H) ; Op 0pz/H

FStanOo + tanOp
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . (25)
FS - tanOotanOp

and KMZ = tan 2 (45° - 0MZ/2) ........................... (26a)

--------------------------- ........................ (26b)

520

(FS.tanBo + tanBp)
where 0MZ tan -1 ( - - - - - - - - - - - - - - - - ) ................ . (27)
(FS - tanBotanBp )

Substituting Equation 25 in Equation 24 permits the factor of safety for

the soil (FS) to be determined :

i. e. FS -------------------------- ................. (28a)

Note that as FS also occurs in the term KMZ, Equation 28a must be solved
iteratively. More rapid convergence is obtained by casting the equation
into the following form for iteration purposes :

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . . . . . . . . . . . . . . (2 8b)

Proceeding as previously for the assessment of local stability

against rupture enables the following equation to be derived
corresponding to Equation 14 :

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (29)

The values of KMZ are obtained from Equation 26 while FS is derived from
Equation 28 based on the assessment of local bond stability. The
variation of FR and FS over the depth of the structure for a specified
value of F equal to unity is shown plotted against depth factor in
Figure 5. Although the limiting condition corresponding to unit factor
of safety has been specified, this only occurs at the base where
sufficient movement has occurred. Figure 5 shows that the strength of
the soil towards the upper part of the structure is not fully mobilised
by virtue of the fact that FS exceeds unity and also as FR is less than
unity. These results may be contrasted with the corresponding condition
for rotation about the base when the factors of safety are the same. The
influence of wall movement on mobilised friction angle may be seen more
clearly in Figure 6 where the relations between depth factor and
mobiiised friction angle are presented for peak angles of 30°, 40° and
50°.
 521

2.00 r-----------------------------"I

1.60

~
~
~
....0 Fs

~
1.20
B
~"
-0
~
~ 0.80 FR
--
::J

co"
u

0.40 Local stability assessment:

rotation about top

0.00 L-_ _-L_ _.....JL-_ _..L..._ _---L_ _ _L -_ _......._ _---IL..-_ _...

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
Depth factor (l/H)

Fig. 5 Relations between depth factor and F(rupture) for F(spec) of unity

50

r- c\>'P 50°

40 r -
-
'"c
C>
co
c
0
30 r
.;;

~ t--
-0
.~
:0
20 r
0
:2:

10 - l Local stability assessment:

rotation about top
J
O~___~~___~~__~~____~I____~I____~I____~i ____~
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
Depth factor (l/H)

Fig. 6 Relations between depth factor and mobilised friction angle

522

3.2 Overall adherence resistance.

To simplify the analysis the assessment of overall adherence

resistance for outward rotation about the top is based on the average
values of safety factor of the soil, mobilised friction angle and earth
pressure coefficient determined from the local bond analysis described
in Section 3.1. The equation for overall adherence stability on this
basis is as follows :

FOVER Ka I L/H - JKMA/~

F
~~;~~~i~---J~:i;] ........................ (30)
The results obtained from the analysis based on Equation 30 are. shown in
Figure 7 plotted against the average safety factor (FAS) of the soil.
Also shown on the figure is the relation between average lateral
pressure coefficient (KMA) and FAS' A comparison of the results produced
by Equations 21 and 30 corresponding to rotation about the base and top
respectively is presented in Figure 8 in terms of the ratio of FOVER/F
versus FAS' The close agreement between the two curves confirms the
point made at the outset that the total force for both cases is the same
and is assumed to correspond to ~~H2.

4. INTERNAL STABILITY CONSIDERATIONS WITH DIRECT ASSESSMENT OF

STRAINS.

In the previous sections the influence of strain behaviour of the

soil and reinforcement was assessed indirectly hy assuming that the
strains developed would be inversely proportional to factor of safety.
An alternative approach will now be described in which the lateral
strains are evaluated directly, ' thus allowing the mobilised friction
angle to be calculated. The lateral pressures and tension in the
reinforcement may then be determined on this basis. The method is
essentially an extension of the conventional design whereby an
evalua.tion of the strain permits the use of more realistic angles of
friction in the design equations.

4.1 Basis of the method.

The method requires a knowledge of the relations between tensile

load and strain for the reinforcement and between mobilised friction
angle and lateral strain for the soil. The simplest approach is to
assume a linear relation for the reinforcement and a bilineal relation
for the soil (Figure 9).
In this latter case the ultimate friction angle (0CV) is used in
preference to the peak friction angle as potential post-peak reduction
in strength and rapid progressive failure are then avoided.
The relation between strain in the reinforcement (€R) and
tensile load (T) is given by :
 523

1.00 r----------------------------,

0.80

uw

-
~ -.....
0.60 I-
~,.w ...... ~

.3
LL
a
a
.;::;
0.40 t-
r-
--
l=--------------------------~::~~::~-~==-:~::-:~~~ - __
......... Fover
a:'" F

0.20 t-
l Overall stability assessment:
rotation about top
J
0.00 1
L-.._ _-L. 1
__---I_ _ _J I
..I..-_ _--'-_ I
_ _.L..-_ I _ _ _'I - - _ - - - '
_- L

1.12 1.20 1.28 1.36 1.44 1.52 1.60 1.68 1.76

Factor of safety of soil

Fig. 7 Relations between F(soil) and ratio of F(over)/F(spec)

1.00

0.80

uw Top rotation
~
0.60
~,.w
.3
LL
....a
a 0.40
.;::;
a:'"

0.20 Overall stability assessment:

rotation about toe and top

0.00
1.00 1.08 1.16 1.24 1.32 1.40 1.48 1.56 1.64
Safety factor of the soil

Fig. 8 Relations between F(soil) and ratio of F(over)/F(spec)

524

40
Plain strain tests on
medium dense sand

32
Q)

~
cit (ev)

::s'"
Q)

Q)
C> 24
c:
'"o
c:
.;:;
u
:E
"0
16
.~
:.0
o
:2:

o 2 4 6 8 10 12 14 16
Lateral strain (per cent)

Fig. 9 Relations between lateral strain and mobilised friction angle

100

Tension distributions based

on strain analysis
80

c: 60
o
.;;;
c: ' Rigid'Reinforcement
~
"0
~
~
::>
40
u
"iii
U

20

o ~----~----~----~----~--__~____~~____~__~
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
Distance from facing (X/ U

Fig. 10 Tension distributions for 'rigid' and 'flexible' reinforcement

 525

T = m. €R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (31)

The relation between lateral strain in the soil (€S) and mobilised
friction angle is given by

€S < €Cv ........................... (32a)

€S ~ €Cv .......................... (32b)

Where €cv is as shown in Figure 9 and corresponds to the minimum

lateral soil strain to mobilise 0CV. Now for purposes of evaluating
local stability the lateral force (DM) is given by :

The lateral force has to be resisted by the tensile force (T) in

the reinforcement

i.e. T KMl'ZSVSH .................................... (33)

Now for compatibility, the (average) strain in the reinforcement and soil
must be equal :

i.e. € = ------------------------ ................•.. (34)

m[l + sin(n€ + 0KO)]

It is apparent for consistency of the results that the argument of the

sin terms in Equation 34 must not exceed the design value of friction
angle. The above equation has been re-arranged to the following form to
allow a more rapid iterative solution to be obtained :

€i+l ----------------------------- ............... (35)

The strain determined from Equation 35 is then used to evaluate 0 m from

Equation 32. It should be noted, however, that the values obtained
relate to the conditions directly behind the facing and it is of
interest to consider how the strains may vary along the reinforcement.
For metallic or other types of high modulus reinforcement, the
magnitude of the strains will be negligible and any outward displacement
of the facing will produce a corresponding movement at every point on
the reinforcement. Thus the application of a pull-out force would result
in a uniform shear strain at the interface between the soil and
reinforcement. Such a distribution of shear strain would produce a
constant mobilised friction angle along the length of the reinforcement
together with a linear variation in tension from a maximum value at the
526

face to zero at the rear.

In contrast, the application of a pull-out force to an extensible
reinforcement would induce the greatest strain at the face where the
maximum tension occurs. If it is assumed that the strain varies linearly
from the face, to zero strain at a distance L from the face where KO
conditions apply, then the tension developed in the reinforcement at
distance x can be obtained from the following differential equation :

dT

dx

Where ~M = f(x)
where A = surface area of reinforcement
where Cl = constant of proportionality

Taking ~M = a. tan0M = a. tan(0KO + n€) where a = an adherence factor

between zero and unity.

Now € is assumed to vary linearly from €M at the face as determined from

Equation 35 to 0 at distance L

00)/n if 0CV is used in design

dT

dx

2AavaLnCl
Tx ---------.In.[cos(0KO + n€M(l-x/L»)] + C2
n€M

when x=O ; TX = TM

when x = L TX = 0

o=

_____ !~:!~I_~?~~KO_l_
In[-~?~~~KO-:-~~~~-]
[ cos0KO ]

TX i~-I~~~~~~o~~~~~I.ln.~~?~i~K~o!(~~~~!:~i~~2~ ...... (36)

. [cos0KO]
 527

Although the tension distribution produced by Equation 36 is not

considered representative of an actual situation, it allows a comparison
to be made between the calculated results for 'rigid' and 'flexible'
reinforcement (Figure 10) . As could be expected the "rigid"
reinforcement has attracted greater tensile forces because the amount of
strain of the soil is reduced with a corresponding reduction in
mobilised friction. Moreover ,because of the uniformity of the shear
strains at the interface, the tension distribution for the "rigid"
reinforcement varies linearly over the reinforcement. In contrast, the
"flexible" reinforcement shows a non-linear distribution because the
soil strains reduce from a maximum value at the facing to zero at the
rear.
An assessment of local bond stability requires that, for flexible
reinforcement, the lateral strain distribution along the length of the
reinforcement must be known or assumed. Making the same assumption as
before that the lateral strain reduces linearly from the facing to zero
at a distance L allows the resisting force (RM) against pull-out to be
determined :

where ~M = a.tan0M = a.tan(00 + n€)

For the simple case assumed, € varies linearly from the maximum value em
at the facing to zero at L :

i.e. € = €M(l - x/L)

Thus RM is determined from

RM ~ (2B~za).tan{00 + n€M(l - x/L)}dx

o
2B~zaLD (COS00 )
- - - - - - - - -In. ( - - - - - - - - - - - - -) ........................ (37)
n€M (COS(00 + n€M»

The disturbing force corresponds to the lateral force acting over the
area of facing supported by a single element :

Hence the actual factor of safety is given by

528

Now the specified factor of safety is equal to

R
F
D

Ka (COS00 )
i.e - - - - - - - - - - - - - ln ( - - - - - - - - - - - - - - ) . .. . ............. (38)
F KM·n·€M·tan0L (COS(00 + n.€m»

The calculated factor of safety is plotted versus the depth below the
surface in Figure 11 for peak friction angles of 30°, 40° and '50°. The
reasons for the relatively small factors of safety are as follows :
(1) The mobilised friction angle is less than the limiting value
behind the facing so that the lateral force is greater than the minimum
value corresponding to active earth pressure conditions.
(2) The magnitude of the lateral strain redu~es along the length
of the reinforcement with a corresponding reduction in interface
friction so that less pull-out force is developed.
However, as the limiting conditions approach, greater strains will be
induced and the available soil strength will increase until it is fully
mobilised. The development of such strains will rarely be uniform and
the maximum strength may be achieved at some locations while at others
only a small proportion will have been attained. Thus with increasing
strain the former locations will tend to reduce in strength as post-peak
conditions are developed while elsewhere the strength may be approaching
or at the peak value. It is unlikely, therefore, that with extensible
reinforcement the peak strength will be attained simultaneously at all
locations and although the available resistance in adherence will
improve with increasing strain, it will be usually less than calculated
on the basis of ideal limiting strength values. A prudent approach to
design would thus be to assume that the maximum available friction angle
of the soil corresponds to the ultimate friction angle of the soil 0CV .
• Notwithstanding that the ultimate pull-out resistance is likely
to be less than calculated on the basis of fully mobilised strengths,
the local adherence stability based on conventional design equations
will generally be adequate although the factor of safety at the working
condition will be usually smaller than specified. Of greater
significance will be the effect of the smaller mobilised soil strengths
in relation to potential rupture.
The assessment of local stability against tensile rupture was
based on the same method as described in Section 2.1 whereby the lateral
forces and mobilised friction angles determined for adherence stability
are again employed. It should be noted, however, that such an approach
is not necessary in this case as an independent calculation of local
stability against rupture would produce the same strains and mobilised
frictions and thus highlights the much greater consistency of a method
based on a strain assessment. The relations between depth factor and
calculated factor of safety, for the same values of peak friction angle
 529

1.00 r - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

0.80

~
.:::
:Jl
'0 0.60
2
~"
-0
~ 0.40
::l

ro"
u
Local stability assessment:
0.20 strain analysis - bond

0.00 L-_ _-L._ _ _.1.-_ _-..L_ _ _- ' -_ _-..L_ _ _- ' -_ _ _' - -_ _......

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60

Depth factor (Z/H)
Fig. 11 Relations between depth and calculated factor of safety

1.00

0.80

~
.:::
:Jl
'0 0.60
(;
1:)
~
-0
2l
~ 0.40
ro"
u

0.20 Local stability assessment:

strain analysis - rupture

0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
Depth factor (Z/H)
Fig. 12 Relations between depth and calculated factor of safety
530

as used previously, are shown in Figure 12.

The description of a possible design method based on the
assessment of strain has been limited to considerations of local
stability and for a very simple assumption concerning the distribution
of internal strains. The extension of the method to overall stability
would not appear to present any special difficulties and could be based
on a similar approach to that described in Section 3.2 for the case of
rotation about the top of the wall when average values were employed.
Alternatively, in preference to the development of an overall
design equation involving the use of average values, it may prove more
convenient to determine the total resistance by summation of the
contributions made by individual elements incorporating the actual
properties of the soil and reinforcement at each level. As far as the
use of alternative patterns of strain behaviour is concerned, the main
problem is that the basis for more complex assumptions may prove
difficult to justify and will lead to more involved design equations.
However,the latter difficulty may be overcome by the use of Southwell-
type plots of the relations between tensile load and extension for a
geotextile and between mobilised friction angle and lateral strain for
soil which generally allow both forms of relation to be linearised
(Murray, 1980) .

5.0 EXTERNAL STABILITY CONSIDERATIONS.

The assessment of the external stability of an earth retaining

structure is carried out to ensure that the wall does not fail by
sliding or overturning and, when cohesive soils are present pehind or
beneath the wall, to ensure that the wall is safe against a potential
slip failure. This latter mode of possible failure is usually assessed
by the Slices Method (Bishop,1955) , or similar, and involves an
iterative procedure for determining factor of safety and mobilised shear
strength. Such an approach is consistent with the methods described
earlier for internal stability and will not be considered further. The
following sections describe procedures which attempt to also bring
consistency into both sliding and overturning aspects of external
stability.

5.1 Resistance to sliding.

The safety factor against forward sliding of the wall on the basis of
limiting strengths is given by :

W.tan0p
F .............................................. (39)

The conditions for limiting equilibrium are

and the limiting width of wall (LL) at the point of sliding is obtained
from :
 531

Ka· z
............................................. (40)
2tan0p

Thus the design length (Ln) is related to the limiting length by a

factor of safety

Ln = F.LL .................................................. (41)

As assumed previously, up to a condition of zero strain, the greater

the factor of safety against sliding the more the soil will be
constrained such that less of its shear strength may be mobilised. Thus
the actual safety factor against sliding is given by :

FSLIn €S>0 .............................. (42a)

and for equilibrium Ln

i.e.
2tan0p

Assuming tan0M = tan0p/FS

then FS = F.Ka/KM ...... ; ..................................... (43)

where KM -----------------------

Thus Equation 43 can be solved iteratively for FS .

To preserve the equality of Equation 42a the factor of safety
against sliding will correspond to unity, until the zero strain
condition is reached. An anomaly occurs, however, as to how the equality
of Equation 42a is preserved when the specified factor of safety is
increased beyond the state when zero strain is attained. One possible
approach is to assume that although KM remains constant at KO, the value
of tan0M in the numerator of Equation 42a reduces to compensate for any
further increase in the length Ln. This assumption implies that FS will
further increase to reduce the mobilised sliding friction over a greater
length. The equation for the factor of safety against sliding is
therefore given by:
532

FSLID o .............................. (42b)

Clearly a stage may be attained with large values of specified

safety factor when further increases in LD will not produce any
redistribution of the applied sliding force and any additional length
will be redundant. At this stage the redundant length will need to be
considered as a reserve of safety factor which does not contribute to
the equilibrium of Equations 42.
In preference to the specified safety factor F which assumes
fully mobilised strengths, or to the value of safety factor against
sliding FSLID which generally corresponds to unity, a more realistic
measure of the safety factor against sliding is provided by the safety
factor of the soil FS' The relations between FSLID based onoEqu~tions 4~
and specified safety factor for peak friction angles of 30 ,40 and 50
are shown in Figure 13.

5.2 Stability against overturning or bearing failure.

The conventional method of designing against oVeTturning assumes

that the wall behaves as a rigid body when resisting the thrust from the
backfill. It is further assumed that a trapezoidal pressure distribution
will be developed at the base of the wall. The method is essentially a
permissable stress approach in which the width of the wall is determined
by an allowable bearing capacity for the foundations. The limiting
strength of the backfill is employed in determining the earth pressure
coefficients and a usual requirement is to avoid the development of
tensile soil stresses in the foundations. The equation for the maximum
foundation pressure on this basis is given by :

af = ~H(l + Ka .H 2/L 2 ) ...................................... (44)

where af must not exceed the allowable bearing pressure qa'

Thus one definition of safety factor against overturning may be
given.by :
qL
F .................................................. (45)
af

where qL is the ultimate bearing capacity of the soil.

As was the case for internal stability it is apparent that there are
inconsistencies in such an approach as the greater the ultimate bearing
capacity relative to the applied pressure , the smaller the amount of
yielding taking place with less scope for mobilising the shear strength
of the backfill. An alternative method is described which involves an
assumption that the mobilised shear strength is reduced in inverse
proportion to the factor of safety against bearing failure. The
procedure is therefore similar to that employed previously for
 533

2.50
External stability:
sliding resistance assessment

2.00

~
~
~
.....a 1.50
B0
~
-0
j!l
'" 1.00
:;
0
"iii
II

0.50

0.00 L -_ _- ' -_ _ _- ' -_ _---L_ _ _...L.._ _----'_ _ _.....L._ _ _"'--_ _.....

1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00

Specified factor of safety

Fig. 13 Relations between specified and calculated factor of safety

1.00 .-=----------------------------,
cpp = 50° cjl'p = 40° cjl' = 30°
........ '=----.:.:..-=.:--==-~-..:.:~- - - ~~
0.80

uCl)

~ 0.60
~
E
a
E-
LL
.....a 0.40
a
'';:;

0:'"
0.20 External stability:
overturning moment assessment

0.00 L - - _ - ' -_ _......JL--_ _...L.._ _---1_ _ _.l.-_ _--L.._ _ _1....._ _..J

1.00· 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60
Safety factor of soil

Fig. 14 Relations between F(soil) and ratio of F(mom)/F(spec)

534

evaluating the internal stability characteristics of reinforced soil

walls.
From Equations 44 and 45 the limiting bearing capacity of the soil is
given by :

qL ~ F-yH(l + Ka .H 2 /L 2 ) ..................................... (46)

Now if the shear strength of the soil is not fully mobilised the lateral
earth pressure coefficient increases to KM and the actual factor of
safety against overturning (FMOM) is given by :

qL ~ FMOH·-yH(l + KM·H 2 /L 2 ) ................................. (47)

Equations 46 and 47 may be solved for FMOH provided the factor of safety
of the soil (FS) is known :

i.e .................................. (48)

F

One technique is to employ the same value of FS as was determined in the

assessment of stability against sliding described in Section 5.1.
Assuming that tan0M tan0p/FS, then the lateral earth pressure
coefficient KM can be determined employing the same equation as given
previously allowing Equation 48 to be solved. The relations between the
ratio of FHOM/F determined from Equation 48 are shown in Figure 14
plotted versus the safety factor of the soil FS'
It should be noted that friction between the backfill and rear of
the wall has been ignored in the analysis . In some cases where the
backfill is attempting to move down relative to the wall this will
result in an underestimate of the factor of safety. However, it seems
better to ignore "wall friction" unless reliable data on the possible
movements are available as in other cases the wall may be attempting to
move down relative to the backfill with an increased overturning effect
and reduction in factor of safety. Thus the use of zero wall friction
prov~des a reasonable compromise between these two extreme situations.

6. FURTHER ASPECTS OF SAFETY FACTOR

The considerations of safety factor have been limited, so far, to

the influence of strain characteristics on behaviour and how the
calculation procedures may be modified to take account of such behaviour
in design. It is apparent from the foregoing analysis that the with
greater constraints on soil strain larger forces have to be carried by
the structure. Thus an important factor in reducing costs is the
development of techniques which avoid inhibiting soil strains but still
ensure stable and serviceable structures. Hethods of achieving this have
be~n recently proposed and are the subject of ongoing research (HcGown
et al,1987).
There are a number of other factors influencing the performance
of reinforced soil structures which are normally accounted for by
selecting an appropriate factor of safety. In particular there is the
 535

variability of material properties, construction and design tolerances,

the influence of construction and deviations in the applied loads. The
application of an overall factor of safety to account for these
variables for structures reinforced with polymeric materials may prove
uneconomic in view of the wide range of strength and strain
characteristics of these materials. Partial factors of safety have been
proposed, therefore, to enable the influence of each factor to be
considered separately (Murray and McGown,1987).
The partial factors to be considered in the design of reinforced
soil structures are listed in Table 1. The values 0 ml and 0 m2 must be
determined for the specific material in the particular application. The
recommended procedure is to establish the variation in properties
between control specimens and those obtained from strength tests on
normal production material in controlled laboratory conditions tested
under both constant rate of strain and sustained loading in both ·the
short and long term. The partial factors 0 m2 is also best evaluated from
such strength tests and comparisons with control specimens to enable the
influence of site damage and environmental effects to be assessed.
Figure 15 shows schematically how the load versus strain relation of
polymeric reinforcement may be influenced by site damage effects.
The partial factors ~fL and ~f3 are normally provided in Codes of
Practice relating to the more general aspects of civil engineering works
(e.g. B.S. 5400,1980).

TABLE 1

PARTIAL FACTORS CONTRIBUTING TO THE OVERALL FACTOR OF SAFETY

PARTIAL FACTOR PURPOSE

To cover the possible reductions in

~Ml material properties compared to the
the properties of control specimens.

To cover for site damage and

construction or manufacturing tolerance
~M2 on site, such as misalignment,
undulations and mis-shaped products.

~fL To cover unfavourable deviations in

loads.

~f3 To cover errors or inaccuracies in the

design method.

Overall factor of safety (~Ml * ~M2 * ~fL * ~f3)

536

--
Undamaged
)( reinforcement

0,-
. ---._)( Site damaged
reinforcement

""0
'"o
-'

Strain

Fig. 15 Influence of site damage on load versus strain relation of polymer

reinforcement shown schematically
 537

7. CONCLUSIONS.

l. In the design of reinforced soil structures it is normally assumed

that the full strength of the soil is mobilised and that the properties
are unaffected by the presence of the reinforcement or boundary
constraints. It has been demonstrated that such assumed behaviour is
unrealistic and tends to inhibit progress towards a better understanding
of how reinforced soil structures actually behave.

2. An alternative method of design has been described in which the

mobilised soil strength is related to factor of safety. This method
highlights the fact that with increasing support by other components in
the reinforced soil system, e.g. reinforcement, facing and foundation,
the work done by the soil reduces. In effect the soil inadvertently
develops a "safety factor".

3. The alternative approach has been shown to be applicable to design in

terms of both internal and external stability and appropriate equations
have been developed for the normal assumption of outward rotation about
the toe as well as for outward rotation about the top which is
considered more relevant for designs involving part-height facing
panels. Although the method does not truly represent strain behaviour,
it provides much greater consistency and compatibility in design while
at the same time preserving the simplicity of the existing methods and
would seem to be particularly suited for use with "stiff"
reinforcements.

4. The increasing use of geotextile reinforcements has highlighted the

need for an alternative design approach which considers strain
behaviour. One such method has been described and the equations for
internal local stability presented. The analysis has been limited to the
simple case of linear variation of strain in the soil and reinforcement
but could readily be extended to more complex situations.

5. Because of the wide range of strength and strain characteristics of

polymeric reinforcements, the most economic approach in the design of
reinforced sbil structures employing such materials is likely to involve
the use of partial safety factors.

8. ACKNOWLEDGEMENTS.

The work described in this paper forms part of the research programme
of the Transport and Road Research Laboratory and the paper is published
by permission of the Director. The author is particularly grateful to
Mr. I. F. Symons of TRRL for a number of helpful suggestions in the
preparation of this paper.
538

9. REFERENCES.

Bishop,A.W. (1955) The use of the slip circle in the stability analysis
of slopes.Geotechnique 5,pp 8 - 17 ,London

British Standards Institution (1980).BS5400 Steel, concrete and

composite bridges. British Standards Institution, London.

Burland,J.B. Potts,D.M. and Walsh,N.M. (1981). The overall stability of

free and propped embedded cantilever retaining walls. Ground
Engineering, Vol. 14, No. 5,London

Department of Transport (1978) Reinforced earth retaining walls and

bridge abutments. Technical Memorandum (Bridges) BE3/78.London

Harr ,M (1966). Foundations of theoretical soil mechanics. McGraw-Hill

Book Company Ltd., London

McGown,A ,Murray,R.T. and Andrawes,K.Z.(1987) Influence of wall yielding

on lateral stresses in unreinforced and reinforced fills. Dept. of
Transport, TRRL Research Report 113 ,Crowthorne ( Transport and Road
Research Laboratory).

Murray,R.T. (1980)Fabric reinforced earth walls: development of design

equations. Ground Engineering, October,pp 29 - 38 ,London

Murray,R.T. and McGown,A (In Press).Assessment of time dependent

behaviour of geotextiles for reinforced soil applications. Proc. Seminar
on Long-Term Behaviour of Geotextiles, St-Remy-les-Chevreuse, 4 6
November 1986, ITBTP, Paris.

Symons,I.F. (1983). Assessing the stability of a propped in-situ wall in

overconsolidated clay. Proc. Instn. Civil Engnrs., Part 2.

Crown Copyright. Any views expressed in this paper/article are not

necessarily those of the Department of Transport. Extracts from the text
may be reproduced, except for commercial purposes, provided the source
is acknowledged.
 539

LIST OF SYMBOLS.

A Surface area of reinforcement

a Cross-sectional area of reinforcement
B Width of reinforcement
Constant of proportionality
Constant of integration
Disturbing force applied to a single element based on Ka
conditions
DM - Disturbing force applied to a single element base on KM
conditions
F Specified factor of safety
FA - Factor of safety against local adherence
FAS - Average safety factor of the soil over the depth of wall
FR - Factor of safety against local rupture
FS - Factor of safety of the soil
FMOM - Factor of safety against overturning or bearing failure
FSLID - Factor of safety against sliding failure
FSi,FSi+l - Factor of safety of soil after i,i+l iterations
H Height of reinforced soil wall
i Number of reinforcing element counting from top of structure
Counter for number of iterations
Ka - Coefficient of active earth pressure
Kma - Average mobilised coefficient of earth pressure over the
height of the wall
- Coefficient of active earth pressure variable with depth
Coefficient of earth pressure when strength of soil is not
fully mobilised
KMZ - Coefficient of earth pressure at depth z when strength of soil
is not fully mobilised
KO - Coefficient of earth pressure "at-rest"
L Length of reinforcement
Design length of reinforcement
Limiting length of reinforcement
m Coefficient defining the slope of the load versus strain of
the reinforcement
n Coef~icient defining the slope of the mobilised friction
angle versus lateral soil strain relation up to a strain of
€CV
N Number of reinforcements in a vertical column
P Total force based on Coulomb analysis
PL - Limiting stress in a reinforcement
PM - Permissable stress in a reinforcement
qL - Limiting bearing pressure in foundations
Allowable bearing pressure in foundations
Resisting force developed by a single element at limiting
conditions
Resisting force developed by a single element at mobilised
strength conditions
Horizontal spacing of reinforcing elements
Vertical spacing of reinforcing elements
Tension in reinforcement
Tension to be resisted by the i th element
Maximum tension in a reinforcement
540

TT - Total tension to be resisted by reinforcements

TX - Tension in reinforcement at distance x from facing
TZ - Tension to be resisted at depth z
W \,Teight of reinforced soil region
x Distance from facing to point on reinforcement
z Depth from top of structure to element under consideration
Coefficient expressing interface friction angle as a
'" proportion of the soil friction angle
{3 Orientation of failure plane from the vertical
'Y Unit weight of soil
'YMl - Partial factor to take account of variations in material
properties
'YM2 - Partial factor to take account of site damage and
environmental effects
'YfL - PartiaL factor to take account of deviations in applied
loading
'Yn - Partial factor to take account of construction tolerances,
connection misalignment and associated effects
8L Projected length of reinforcement beyond failure surface
Strain in the reinforcement
Strain in reinforcement based on load versus extension test
Minimum lateral strain to mobilise 0CV
Strain determined after ith iteration
Maximum strain in reinforcement
Lateral strain in soil
Slope of the mobilised friction angle versus depth relation
for facing rotating about the top
0CV - Ultimate friction angle of the soil
0KO - Friction angle corresponding to "at-rest" conditions
0M Mobilised friction angle of the soil
0MZ - Mobilised friction angle of the soil at depth z
0p Peak friction angle of the soil
0pZ - Peak friction angle that can be mobilised at depth z
0Ji- Interface friction angle
Ji-L Limiting interface friction coefficient
Ji-M Mobilised interface friction coefficient
af Vertical pressure beneath front of wall
ah Horizontal earth pressure
av Vertical earth pressure
BO Initial friction angle at depth z (less than or equal to
0KO ) when wall is rotated about top
Bp Maximum friction angle at depth z
APPLICATION OF FINITE ELil1ENT TECHNIQUES TO THE ANALYSIS OF
REINFORCED SOIL WALLS

R.K. ROWE and S.K. HO

GEOTECHNICAL RESEARCH CENTRE, UNIVERSITY OF HESTERN ONTARIO,

LONDON, CANADA.

1. INTRODUCTION
The beneficial effect of incorporating tensile inclusions
within a soil mass is well recognized and has been demonstrated
by the successful construction of numerous reinforced soil
walls using facings ranging from relatively rigid full face
concrete panels to a flexible geotextile "skin", and reinforce-
ment ranging from relatively stiff steel strips or meshes to
geotextile sheets of low stiffness. Design methods for rein-
forced-soil walls ("7hich are discussed in detail in other
papers at this workshop) are typically based on limit equili-
brium calculations "Thich do not explicitly consider deforma-
tions or interaction beh7een the inclusion and the soil.
In the case of Reinforced Earth (R) walls, there is now
twenty years of empirical evidence to suggest that the approxi-
mate method of analysis, in conjunction with the normally
adopted soil properties and safety factors, provides a safe de-
sign which gives acceptably small deformations under working
conditions. This may also be the case for some geotextile and
geogrid reinforced walls designed using current practice, how-
ever the "Tide range of facings, backfill materials and proper-
ties of the reinforcement leads one to question the generality
of the simplified methods of analysis that are being proposed
for the design of reinforced soil walls with geosynthetic in-
clusions. Questions that can be raised include:
- what is the effect of reinforcement extensibility;
- what is the effect of using different facings and construc-
tion techniques;
- what is the-effect of soil-facing-reinforcement interaction
both during construction and subsequently (over the life of
the structure);
- under what circumstances might one expect strain-softening
within the soil mass and what influence do the properties
of the reinforcement have on strain softening;
- what is the "Factor of Safety" of a reinforced soil wall?
In prinClple, these questions could all be answered by the
construction and monitoring of a large number of full scale
field t~st walls. Unfortunately, the cost of performing and
adequately monitoring a sufficiently large number of full sc~le
walls is so large that it is not practical to perform a detail-
ed experimental study. "Numerical experiments" or simulations
provide an alternative and more cost effective means of perform-
ing such a study.
Finite Element techniques have the potential to allow us to:
541

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 541-553.
© 1988 by Kluwer Academic Publishers.
542

- improve our understanding of observed behaviour in field

trials;
- model the complete response of a reinforced soil wall up to
collapse;
- examine the effects of changes in the elements of the
system (i.e. the properties of the reinforcement, the soil
or the facing);
- investigate changes in construction procedures and the
nature of the system; and
- study the interaction between the foundation characteris-
tics and the performance of the reinforced soil wall (i.e.
to evaluate the effect of local yield and settlement within
the foundation for reinforced walls on less than ideal
foundations).
The Finite Element technique is well recognized as being a
very powerful tool and examples of its application to modelling
reinforced soil walls (especially Reinforced Earth (R) Walls)
can be found in the literature (eg. see ASCE Symposium on Earth
Reinforcement, 1978). However, the use of the technique for
performing studies which could answer the questions raised
above is subject to some important constraints.
Firstly, if the technique is to provide answers to some fun-
damental questions (as opposed, say, to simply calculating the
response of a structure under working conditions), then the
formulation must provide a good description of all components
of the reinforced soil wall system.
Secondly, a numerical study such as this should not be per-
formed in isolation. It is essential that the formulation and
technique be validated against (a) limiting analytical bench-
mark solutions; and (b) available data from both model and
field tests. Indeed, the paucity of well documented field
cases involving polymer reinforced ",alls which could be used to
validate finite element models, has been a key factor constrain-
ing the use of Finite Element technique for the detailed study
of reinforced soil wall behaviour.
Finally, although a numerical study may represent a cost-
effective. alternative to a full scale field study, it is not
without cost. A good non-linear F.E.analysis requires an ex-
perienced "driver", considerable data and data preparation
time, and considerable computer time.
Thus the objective of this paper is to review the past appli-
cation of Finite Element techniques for the analysis of rein-
forced soil walls and to comment on some of the factors re-
quiring consideration if the technique is to be used to answer
some of the questions raised in this section.

2. REPRESENTATION OF A REINFORCED SOIL \VALL SYSTEM

There are two techniques for modelling reinforced soil walls
using finite elements, viz. composite and discrete representa-
tion of 'the constituents. Each approach has its own advantages
and limitations as discussed below.

2.1 Composite representation

Composite formulations are based on an extension of continuum
concepts to a macro level of observation whereby the entire re-
inforced soil mass is treated as an anisotropic, homogeneous
 543

material (eg. Romstad et al., 1976; Chang and Forsyth, 1977).

In most formulations of this approach, the composite element
stiffness is formed by superimposing the stiffness of the rein-
forcement and soil (assuming no slip) and so the distribution
of elements need not be directly related to the distribution of
reinforcement within the soil mass (eg. one or more layers of
reinforcement may be included in a single composite element).
This can result in a substantial decrease in the number of
equations that must be solved for a given physical problem (as
compared to discrete element formulations) with consequent sub-
stantial savings in computer time. Clearly, a disadvantage of
the approach is that it does not provide direct information
concerning the stress and strains at the interface of the rein-
forcement, nor does it provide information concerning localized
deformation near the edge of the reinforced soil mass. ,
As noted above, most composite formulations assume a no-slip
condition between the reinforcement and soil and hence are only
applicable provided that the applied loads and distribution of
reinforcement are such that little or no slip would be expecte~
Attempts have been made to broaden the applicability of the
approach by modelling slip in the composite representation of
reinforced soil. This can be achieved by introducing extra
nodal displacement variables relating the relative displace-
ments betyleen the soil mass and reinforcement (eg. see Herrmann
and Al-Yassin, 1978; Naylor and Richards, 1978). However, the
introduction of the extra variables required to model slip to-
gether with the consequent restrictions on the number of layers
of reinforcement which can be included in a composite element
largely eliminates the computational savings of a composite
approach (i.e. which arise if it is not necessary to represent
every reinforcing element) while still being subject to some of
the disadvantages of the approach previously discussed.

2.2 Discrete representation

In a discrete representation of a reinforced soil system (eg.
see Al-Hussani and Johnson, 1978; Andrawes et al., 1982;
Andrawes ~t al., 1980; Banerjee, 1975; Rowe, 1984 ), the soil
mass, the reinforcement, the facing and the interface between
"structural" element and the soil are all independently repre-
sented by discrete elements. This approach provides direct in-
formation concerning the stresses and deformations at the inter-
face of the reinforcement, the variation in stresses and
strains (within the soil) between layers of reinforcement, and
the localized deformations near the edge of the reinforced soil
mass. However, to obtain accurate results, it is necessary to
ensure that the choice of finite element and the distribution
of finite elements (particularly between layers of reinforce-
ment) provide sufficient "freedom" for realistic stress and
strain distributions to be developed. To capture these de-
tails, a number of bays of elements should be provided between
layers of reinforcement. The actual number of elements re-
quired will depend on the problem and the type of finite ele-
ments being used. An indication of the suitability of a given
finite element mesh can be obtained by detailed comparison of
stresses and strains calculated at various stages in an analy-
sis using this mesh, with corresponding values calculated using
5M

a substantially refined mesh.

In summary, the only disadvantage arising from the use of a
discrete representation of a reinforced soil system is that it
requires more computing resources and effort in preparing input
data than a composite approach. Weighing this disadvantage
against the advantages of potentially more accurately analyzing
the reinforced-soil system, the discrete approach is considered
to be most appropriate for detailed investigation of reinforced
soil wall behaviour. Furthermore, Herrmann and Al-Yassin's
(1978) claims that discrete and composite approaches can be
applied with equal accuracy (which are based on the analysis of
a simple highly idealized problem) should be viewed with consi-
derable caution since it has not been demonstrated that .this is
indeed true for realistic problems which involve construction
simulation, non-linear elastic-plastic behaviour of the soil,
either steel or strip reinforcement and different interface
properties in pullout or direct shear modes of failure.

3. MODELLING OF DISCRETE COMPONENTS

Modelling of the discrete components of a reinforcea soil
system involves consideration of the type of finite element and
the constitutive relationship that will be adopted. Table 1
summarizes the type of element and constitutive model which
have been used in the past by a number of investigators analyz-
ing reinforced soil walls.

3.1 The soil

The literature abounds with continuum elements which could be
used to model the soil. Based on past experience with related
problems, it would appear that many of the available elements
can be used provided that sufficient attention is paid to
checking the adequacy of associated finite element mesh, how-
ever it should also be noted that:
(1) particular care is required in the choice of mesh when using
3 noded (constant strain) triangles;
(2) lower order (eg. 4 noded) quadrilateral elements may give
poo~ results and/or be computationally very inefficient;
(3) higher order quadrilateral elements (eg. 8 noded isoparame-
tric) used in conjunction with reduced integration may give
rise to physically unacceptable results under some circum-
stances.
The choice of element and details of the mesh design are
likely to be far more critical when attempting to predict
collapse than when simply calculating behaviour under working
conditions.
To date, the majority of analyses which have been performed
(see Table 1) have assumed an elastic or non-linear elastic
(hyperbolic) constitutive relationship. Elastic models which
do not consider the variation in soil stiffness with increasing
stress level during construction are of doubtful validity since
the pressure sensitive nature of the. stress-strain characteris-
tics of the backfill may have a significant influence on the
stresses and displacements developed within the reinforced
structure. The simplest way of avoiding this limitation is to
adopt a non-linearity based on Janbu's equation (viz)
TABLE 1. Summary of some finite element analyses of reinforced soil walls

Ref erence------ Type of Soil Model Reinforcement Soil/Reinforce- Construc-

Soil Element Model ment Interface tion Simu-
Model lation*
Al-Hussaini & 5-node '---hyper bo li c linear eTastic- joint element (4 ) &
Johnson, 1978 incompatible plastic bar with hyper- surface
quadrilateral element bolic shear loading
stress-
displacement
relationship
Banerjee, constant strain elastic elastic no-slip (6)
1975 triangle isotropic bar element
Chang & 5-node hyperbolic composite no-slip (3)
Forsyth, 1977 incompatible with soil
quadrilateral
Herrmann & 4=node hyperbolic a )Composi te extra nodal (3 )
AI-Yassin isoparametr ic with soil displacement
quadrilateral b) elastic-plastic variables
beam element between soil
& reinforcement
Naylor, 1978 6-node linear composite (1)
Naylor & quadrilateral elastic with soil
Richards, 1978
Romstad ~t al., S:"-node hyperooTic a )compoS-i te a) no-slip (2)
1976 incompatible with soil b) no-slip
quadrilateral b)beam element
Seed et al., 4-node a) hyperbolic linear elastic normal and (5)
1986 isoparametric b) hysteretic bar element shear
element spring
Shen et al., 5-node hyperbolic composite no-slip (3)
1976 incompatible with soil
quadrilateral
* (I)-Assume initial Ko condition and then remove traction at wall ¥3urface. (2) Turn on
gravity for unpropped system. (3) Construct and turn on gravity layer by layer. (4) As in
(3) but no lateral displacement is allowed at wall face. (5) As in (3) but including consi- Ul
deration of compaction effect. (6) Not reported. ~
546

(E/p ) = K(aIP )n (1 )
a a
where E is the Young's modulus of the soil, a is the minor
principal stress or the mean stress depending on the details of
the formulation, P a is atmospheric pressure and K and n are the
material parameters. This non-linearity is included in the
"hyperbolic" model and can also be readily included in non-
linear elastic-plastic models (eg. Rowe, 1986). It should be
recognized that the modelling of "yield" implicit in Eq. 1 is
only approximate and is not appropriate for situations where
there may be cyclic loading (eg. see Zylynski et al., 1978).
However, there is considerable evidence to suggest that this
approach can provide reasonable results for problems involving
monotonic loading (as is generally the case in modelling wall
construction) .
Non-linear elastic (hyperbolic) models can be expected to
provide acceptable results at low stress levels (eg. when there
is a large "factor of safety"), however since they are based on
elastic theory they can not correctly model plastic failure and
plastic strains within the soil mass (it is noted that the use
of a cohesion- intercept c and friction angle ¢ in a hyperbolic
model does not imply that the model is a plasticity model--eg.
see Duncan, 1980). Numerous plasticity formulations have been
proposed in the literature. The simplest of these involves
Mohr-Coulomb failure surface and a non-associated flow rule.
This model has been successfully applied in the analysis of
geotextile reinforced embankments (eg. Rowe, 1982; Rowe, 1984;
Rowe et al., 1984; Rowe and Soderman, 1984). This form of ana-
lysis can be readily modified to include the consideration of
a non-linear failure envelope commonly encountered with granular
materials (see Rowe et al., 1982). These models can be expected
to model the soil behaviour up to arid including failure. By
examining the results of studies performed using this class of
model it is possible to assess the magnitude of the strains to
be expected prior to collapse of the structure and hence to
make some initial assessment of potential significance of
strain 90ftening. However, this class of model is not suitable
for modelling strain-softening behaviour and indeed the model-
ling of localization and strain softening even for unreinforced
granular materials requires considerable additional research.

3.2 The reinforcement

The reinforcement can be modelled using a one dimensional bar
element. Non-linearity of the stress strain behaviour and
yield can also be readily modelled by making the element stiff-
ness a function of stress (or strain) level. Breakage (snap) of
the reinforcement can also be modelled however this involves the
redistribution of stresses developed in the reinforcement prior
to breaking and erroneous stress distributions can be obtained
unless particular care is taken with the numerical algorithm
used to redistribute these stresses.

3.3 The facing

Depending on the type of facing being considered, it may be
appropriate to use continuum elements, beam elements or bar
elements. The use of continuum elements or beam elements to
 547

model relatively rigid facing is quite straight forward. The

more difficult problem is to correctly model wrap around facings
(i.e. cases where a facing consists simply of the geotextile or
geogrid reinforcement (which has been modelled with bar ele-
ments) being "wrapped around" the soil and "locked" into place
by the overlying backfill). Correctly modelling the stresses
and deformations resulting from the form of construction is not
a trivial exercise.

3.4 The soil-reinforcement interface

The interaction between the soil mass and the reinforcement
can be modelled by introducing soil-reinforcement interface
elements. This can be achieved in a number of ways including
the use of joint elements, nodal-compatibility slip elements or
by substructuring. Common approaches to modelling the soil re-
inforcement interface involve three nodes at each point alQng
the reinforcement; one attached to the soil above the reinforce-
ment, one on the reinforcement, and one to the soil below the
reinforcement. The nodal-compatibility slip element (which may
be formulated initially in terms of normal and tangential
springs with very high stiffnesses) (i) ensures compatible dis-
placement between a pair of dual nodes (one attached to the
soil and one attached to the reinforcement) until a Mohr-
Coulomb failure criterion is reached, and (ii) replaces the com-
patibility conditions by a failure condition and dilatancy
equation once the interface strength is exceeded. Joint ele-
ments allow relative deformation of the soil and reinforcement,
prior to failure of the interface, based on some assumed con-
stitutive relationship of what is in effect an interface layer
between the reinforcement itself and the general soil continuum
(eg. Andrawes et al., 1980, 1982 used a hyperbolic model to re-
present the interface behaviour).
In its simplest form, the joint element may be comprised of a
pair of normal and tangential springs. Clearly, as the stiff-
ness of a joint element increases, it tends to a nodal-compati-
bility slip element and the distinction between the two is re-
lated to .the question of whether a distinct interface layer
exists or whether the deformations at the interface (prior to
failure) are simply due to the interaction between the rein-
forcement and the soil on either side of the interface. If
there is good experimental data indicating that a distinct
interface layer exists with experimentally defined stress-
strain characteristics then this can be readily modelled as a
joint element or as a thin layer of continuum element (with
slip still being modelled using a nodal-compatible slip ele-
ment). In the absence of this data, a nodal-compatibility slip
element would seem appropriate.
Any modelling of .in terface behaviour must consider three
possible mechanisms of failure as noted below.
(a) If there is insufficient anchorage capacity, failure will
occur at the soil reinforcement interface above and below the
reinforcement as the reinforcement is pulled out of the soil.
This "pullout" mode involves displacement of the reinforcement
relative to the soil on both sides of the reinforcement.
(b) If the shear strength of the soil reinforcement is less than
the shear strength of the soil alone, then failure may occur by
548

sliding of the soil along the upper surface of the reinforce-

ment and the upper soil mass moves relative to both the rein-
forcement and the underlying soil.
(c) The soil below the reinforcement (usually the foundation
soil if one has a soft foundation) may be squeezed out from
beneath the lowest reinforcement layer (and the entire rein-
forced soil wall). In this case, the lower soil may move rela-
tive to the reinforcement and the overlying soil.
If the reinforcement is in the form of a sheet, completely
separating the soil above and below the reinforcement, then the
interface resistance can be readily determined by direct shear
tests (see Rowe et al., 1985). In this case, provision forslip
at the interface is the same irrespective of the mechanism of
failure (that is, direct shear or pullout). However, if the
reinforcement takes the form of a geogrid, with openings which
are large compared to the grain size of the soil, or if the re-
inforcement consists of separate reinforcing strips (eg. steel
strips), then special care is required to correctly model the
failure mechanism. For these materials, the interface shear
resistance in direct shear (eg. if there is sliding of the soil
along the upper surface of the reinforcement) may be substan-
tially higher than the interface resistance in pullout (for
example see Rowe et al., 1985). In modelling these materials,
it is necessary for the formulation of the interface element to
be such that it can automatically detect whether it is in a
direct shear or pullout mode and to then select the appropriate
interface parameters to model this mode of shearing. Thus the
behaviour of the interface element on one side of the reinforce-
ment is related to the behaviour of the interface element on
the other side (since the mode of shearing can only be assessed
by consideration of the direction of shear on either side of the
reinforcement) .
For planar reinforcement independent movement of the soil may
occur above and below the reinforcement following either a direct
shear or pullout failure. For strip reinforcement, independent
movement of the soil above and below the plane of reinforcement
can only occur during a direct shear mode of failure. Pullout
of strips is really a three dimensional phenomenon in which the
strips move relative to the soil around them but the soil be-
tween strips remains continuous. As noted by Naylor and
Richards (1978), the common approach of using a conventional
joint element (or nodal compatibility element) implicitly treats
the strips as an equivalent two dimensional sheet and will cause
serious error since it interrupts the transfer of shear stress
through the soil.
Since pullout of strips does represent a truly three dimen-
sional situation, it can only be approximately modelled in a two
dimensional analysis. A number of different approaches can be
adopted. For example, Naylor and Richards (1978) proposed a
compogite formulation which ensured continuity of shear stress
in the soil after pullout by introducing a "conceptual shear
zone". An alternative approach implemented by the authors in
their discrete formulation involves an interface element which
involves a node above the reinforcement, a node on the reinforce-
ment and a node below the reinforcement. Prior to slip, normal
and tangential compatibility between the soil and reinforcement
 549

is enforced by means of very stiff springs. The normal and

shear stresses "above" and "below" the reinforcement are auto-
matically monitored. If a pullout mode of failure occurs (as
inferred by the direction of shear above and below the rein-
forcement together with a l1ohr-Coulomb failure criterion), then
the computer program automatically enforces compatibility bet-
ween the soil nodes "above" and "below" the reinforcement
(thereby maintaining continuous transfer of shear stress in the
soil) while allowing slip between the reinforcement node and
the two soil nodes. The normal force between these nodes is
used to assess the normal forces acting on the strip; the cor-
responding shear resistance (based on a Mohr-Coulomb failure
criterion) between the strip and soil is applied to both the
upper and lower soil node, and as an equilibrating force to the
node on the soil strip. (Since the strip only covers a small
area of the soil, the Mohr-Coulomb parameters must be adjusted
to take account of the actual surface area, per unit width of
wall, which is in contact with the soil.)

4. SH1ULATION OF LATERAL SOIL PRESSURE AND CONSTRUCTION DETAILS

To date, most finite element analyses of reinforced soil wall
systems have employed classical earth pressure theory in simu-
lating the pressure exerted on a reinforced soil wall system.
One approach is to assume that the wall is constructed under
"at rest,K o " conditions, i.e., as if temporary support was pro-
vided to prevent lateral yielding during construction; loading
is then provided by "removing the support" which involves
applying a horizontal traction to the face of the wall equal to
KoyH (eg. see Naylor and Richards, 1978). Another approach is
to assume the reinforced soil wall is constructed in an "active,
Ka" state of failure, with equilibrium being maintained by
applying a traction equal to KayH on the back of the wall (i.e.
assuming the wall is free to translate, or rotate about the top
or bottom; ego see Banerjee, 1975). These approaches are
simple but they also neglect the influence of construction
method and, in general, cannot be expected to provide a good
represeneation of the behaviour of the reinforced soil struc-
ture.
One factor affecting the magnitude and distribution of lateral
pressures, soil movements and reinforcement strains within a
reinforced soil wall system is compaction. Test data and theo-
retical calculations both indicate that the lateral earth pres-
sure due to compaction of the fill behind flexible (and rigid)
retaining walls can be far in excess of values predicted by
classical earth pressure theory (eg. see Seed and Duncan,
1986). In a reinforced soil wall this phenomenon is likely to
have a significant impact on the magnitude of the stresses and
strains in the reinforcement, particularly if compaction equip-
ment is brought close to the face of the wall.
It has also been demonstrated that compaction of the fill can
have an important effect on the changes in horizontal stress
which arise from applied surficial loads behind the wall (see
Duncan and Seed, 1986).
In an attempt to study the effect of compaction on a rein-
forced soil system, Seed et al. (1986) performed finite element
analyses on three case histories. Their analyses employed a
550

hysteretic loading/unloading model in simulating compaction in-

duced lateral pressures. They found that conventional finite
element analyses which do not model compaction induced pres-
sures gave rise to lower estimates of the tensile stress in the
reinforcement compared with analyses which did model compac-
tion; the effect of compaction being most pronounced at shallow
depths. The conclusion from Seed et al. 's study was that field
measurement (in terms of reinforcement force) could be better
predicted when compaction effects are modelled in the analysis.
Another major factor affecting the magnitude and distribution
of lateral pressure exerted by a reinforced soil wall system is
the lateral restraint of the facing. McGown et al. (1987) have
shown that the magnitude and distribution of earth pressures
behind a model reinforced soil wall will depend on both the
stiffness of the wall supports and the number of layers of re-
inforcement. These pressures may range from values correspond-
ing to Ko conditions to values much lower than active earth
pressures. There is also evidence to suggest that the behaviour
of walls with full faced panels will be different to walls with
segmented facings (where each segment can undergo some indepen-
dent rotation and translation) .
It may be concluded that careful finite element modelling of
reinforced soil walls should include simulation of the actual
or expected construction procedure. In particular, modelling
of the facing support (and its removal) may be expected to pro-
vide better results than more conventional approaches which in-
volve applying an assumed pressure distribution (be it Ko or
Ka) to the back of the wall. Perhaps less important, but
nevertheless deserving of consideration, is the simulation of
the effects of compaction. (The importance of compaction will
of course depend on the height of the wall, the type of compac-
tion equipment used and the proximity of compaction equipment
to the face).

5. FOUNDATION-WALL INTERACTION
One of the advantages of a reinforced soil wall over conven-
tional wall systems is that it should be more tolerant of de-
formations and stresses induced by some yielding in the founda-
tion, thereby allowing construction of walls on less than ideal
sites. Unfortunately, conventional methods of analysis (ego
see Gourc et al., 1987) can not provide insight regarding the
effect of foundation movements on the stresses and deformations
of the wall. The finite element method is ideally suited for
modelling the foundation-reinforced soil wall interaction which
would occur when there is yielding in the foundation soils.
Modelling of this interaction will, however, require the use
of a constitutive model that models plastic strains using a
consistent plasticity formulation which can take account of the
influence of rotations in principal stress directions which
will occur near the toe of the wall. As a prerequisite, the
finite element procedure adopted in these calculations should
be capable of accurately predicting bearing capacity collapse
loads and should be calibrated against relevant published bear-
ing capacity solutions (ego Davis and Booker, 1973).
 551

6. TIME EFFECTS
An important concern in the analysis of walls reinforced with
geosynthetics is the effect of the time dependent characteris-
tics of the reinforcing material. In principle, finite element
methods are well suited to modelling creep/relaxation in both
the reinforcement and the soil (the latter being of particular
importance if cohesive backfills are used). Numerous visco-
elastic-plastic finite element formulations have been published
in the literature however, as yet, these techniques have not
been applied to a comprehensive study of reinforced soil wall
systems. Hodelling of this time dependent behaviour is an im-
portant challenge but one that can not be fully met until there
is good quality field (or model scale) time dependent test data
which can be used for validating the finite element calcula-
tions.

7. CONCLUSION
In this paper we have attempted to review the application of
the Finite Element technique to the analysis of reinforced soil
walls. There are still many unanswered questions regarding the
behaviour of reinforced soil structures and the finite element
method provides a very useful tool which can be used to help
anS\oler these questions. However, it has also been emphasized
that there are many different types of finite element analyses
and that considerable care must be exercised in both the selec-
tion of the particular finite element formulation to be adopted
(eg. in the choice of constitutive model for the soil; modelling
of the interface, etc.) and in the detailed application of the
technique (eg. choice of elements, distribution of elements,
construction simulation, etc.). There is also considerable
scope for additional research in developing or adapting techni-
ques for modelling time dependent interaction between the
various components of the reinforced soil system as well as for
developing techniques which model strain softening and locali-
zation (a major problem in itself) within reinforced soil
systems.

ACKNOWLEDGEMENTS
This review forms part of a general programme of research
into reinforced soil and geosynthetics being conducted by the
Geotechnical Research Centre with funding from the Natural
Sciences and Engineering Research Council of Canada under grant
A1007.
552

REFERENCES

1. Al-Hussaini MM, Johnson LD: Numerical Analysis of a Rein-

forced Earth Wall. Proc. ASCE, Symposium on Earth Rein-
forcement, Pittsburg, pp. 98-126, 1978.
2. Andrawes KZ, McGown A, Mashhour MM, IHlson-Fahmy RF: Ten-
sion Resistant Inclusion in Soils. ASCE, J. Geotech. Eng.
Div., Vol. 106, No. GT12, pp. 1313-1326, 1980.
3. Andrawes KZ, McGown A, v7ilson-Fahmy HF, Mashhour MM: The
Finite Element Hethod of Analysis Applied to Soil-Geotex-
tile Systems. Froc., 2nd Int. Conf. on Geotextiles, Las
Vegas, 2, pp. 695-700, 1982.
4. Banerjee PK: Principles of Analysis and Design of Reinforc-
ed Earth Retaining Walls. J. Inst. Highway Eng., 22, pp.
13-18, 1975.
5. Chang JC, Forsyth RF: Finite Element Analysis of Reinforced
Earth Wall. ASCE, J. Geotech. Eng. Div., Vol. 103, No. GT7,
pp. 711-724, 1977.
6. Davis EH, Booker JR: The Effect of Increasing Strength v7ith
Depth On the Bearing Capacity of Clays. Geotechnique, Vol.
23, No.4, pp. , 551- 5 63, 1973.
7. Duncan JM: Hyperbolic Stress-Strain Relationships. Proc.
ASCE, v7orkshops on Limit Equilibrium, Plasticity and Gene-
ralized Stress-Strain in Geotechnical Engineering, pp. 443-
460, 1980.
8. Duncan JM, Seed RB: Compaction-Induced Earth Pressures
Under Ko-Condition. ASCE, J. Geotech. Eng. Div., Vol. 112,
No.1, pp. 1-22, 1986.
9. Gourc JP, Ratel A, Gotteland Ph.: Analysis and Comparison
of Existing Design Methods and Proposal For a New Approach.
Nato Advanced Research Workshop, Application of Polymeric
Reinforcement in Soil Retaining Structures, Royal Military
College of Canada, 1987.
10. Herrmann LR, Al-Yassin Z: Numerical Analysis of Reinforced
Soil Systems. Proc. ASCE Symposium on Earth Reinforcement,
Pittsburg, pp. 428-457, 1978.
11. McGovln A, Andrawes KZ, Murray RT: The Influence of Lateral
Boundary Yielding On the Stresses Exerted by Backfills.

12. Naylor DJ: A Study of r.e. Wall Allowing Strip Slip. Proc.
ASCE Symposium on Earth Reinforcement, Pittsburg, pp. 618-
643, 1978.
13. Naylor DJ, Richards H: Slipping Strip Analysis of Reinforc-
ed Earth. Int. J. for Numerical and Analytical Methods in
Geomechanics, Vol. 2, pp. 343-366, 1978.
14. Romstad KM, Herrmann LR, Shen CK: Integrated Study of Rein-
forced Earth - I. Theoretical Formulation. ASCE, J. Geotech
Eng. Div., Vol. 102, No. GT5, pp. 457-471, 1976.
15. Rowe RK, Lo KY, Tham L: The Analysis of Tunnels and Shafts
in ' Dense (Oil) Sands. Proceedings of the Fourth Inter-
national Conference on Numerical Methods in Geomechanics,
Edmonton, pp. 587 - 596, 1982.
16. Rowe RK: The Analysis of an Embankment Constructed On a
Geotextile. Proc. 2nd Int. Conf. on Geotextiles, Las Vegas,
2, pp. 677-682, 1982.
 553

17. Rowe, RK: Reinforced Embankments: Analysis and Design.

ASCE, J. Geotech. Eng. Div., 110, pp. 231-246, 1984.
18. Rowe RK, MacLean MD, Soderman KL: Analysis of a Geotextile
Reinforced Embankment Constructed on Peat. Can. Geotech.
J., 21, pp. 563 - 5 76, 1984.
19. Rowe RK, Soderman KL: Comparison of Predicted and Observed
Behaviour of Two Test Embankments. Int. J. of Geotextiles
and Geomembranes, 1, pp. 157-174, 1984.
20. Rowe RK: Numerical Modelling of Reinforced Embankments
Constructed on vleak Foundations. 2nd Int. J. on Numerical
Models in Geomechanics, Ghent, pp. 543-551, 1986.
21. Rowe RK, Ho SK, Fisher DG: Determination of Soil-Geotextile
Interface Strength Properties. Proc. 2nd Canadian Symp. on
Geotextiles and Geomembranes, Edmonton, September, 1985.
22. Seed RB, Duncan JM: FE Analysis: Compaction-Induced
Stresses and Deformations. ASCE, J. Geotech. Eng. Div.,
Vol. 112, No.1, pp. 23-43, 1986.
23. Seed RB, Collin JG, Mitchell JK: FEH Analysis of Compacted
Reinforced Soil Walls. 2nd International Symposium on
Numerical Hodels in Geomechanics, Ghent, pp. 553-562, 1986.
24. Shen CK, Romstad KM, Herrmann LR: Integrated Study of Rein-
forced Earth - II: Behaviour and Design. ASCE, J. Geotech.
Eng. Div., Vol. 102, No. GT6, pp. 577-590, 1976.
25. Zylynski M, Randolf MF, Nova R, Worth CP: On Modelling the
Unloading-Reloading Behaviour of Soils. Int. J. Num. and
Analytical Hethods in Geomechanics, 2, pp. 87-93, 1978.
 REVIEW OF SESSION

ANALYTICAL TECHNIQUES AND DESIGN METHODS

The chairman, Rowe, began the session by posing a number of questions that
require answers when considering analysis and design and which the various
speakers would address during the session. They included the following:

1. Should we be trying to estimate deformations using limit equilibrium

analyses?

2. What is the effect of reinforcement extensibility?

3. Is strain-softening of the soils important?

4. How valid is it to find the level of reinforcement required without

considering the interaction between it and the soil?

5. What is the Factor of Safety in our designs?

6. How does one account for construction technique in design?

7. How useful is the finite element method in analysis of reinforced

soil ?

8. Are any of the existing analytical methods providing the needed

answers?

Jewell presented details of his ideas concerning the interpretation of the

soil parameters that should be used in limit equilibrium analysis. They
were based' on simplifications of stress-dilatency concepts. He also
addressed the problem of compatibility of strains between the soil and the
reinforcement. Andrawes questioned how the dilatancy term used could
always be additive to <Pcv. McGown noted the problems created in the
construction process by using "ideal" spacing concepts which produced
variable spacings between reinforcement layers. Leflaive questioned
whether the important point of anisotropic strain was being considered and
also cautioned that one must be careful to differentiate between the
various polymers available when considering the effects of time and
temperature on the stress-strain properties. McGown stated that although
difficult it would be useful to attempt to estimate horizontal strains
from laboratory tests.

Bonaparte then considered the effect of variation in reinforcement

extensibility on the analytical and design processes. He considered the
problem from the perspective of a design engineer who had reasonable
analytical tools for stability analysis but no simple methods for predict-
ing movements. His approach used stress path analys is for the proposed
structure leading to estimates of soil strains, especially the horizontal

555

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 555-556.
© 1988 by Kluwer Academic Publishers.
556

strains. These could by assuming compatibility be related to strains in

the reinforcement. At one point he indicated that the stiffnesses of most
of the common geosynthetics used in North America may limit their use to
walls of about 9 m in height if deformations in the structure need to be
limited to reasonable values of around 1 inch. Jones stated that there
are stiffer geosynthetics in use in Europe that allow considerably higher
walls to be built without excessive deformations. Bonaparte concluded his
presentation by saying that he believed that soils were usually in an
active state immediately behind reinforced soil walls and that limit state
methods were acceptable for simple design. He also suggested that
research was needed on the influence of construction methods and facings
on wall behaviour. on the development of horizontal strains with changing
lateral earth pressures. on the influence of vertical spacing between
reinforcement layers relative to the particle size of the soil and on full
scale trials.

Murray's contribution dealt with the use of Factors of Safety in design

and their implications in an analytical and practical sense concerning the
mobilization of resistance in both the soil and the reinforcements.
McGown pointed out that in certain of Murray's analytical developments the
facing was assumed to be rigid whereas in practice there may commonly be a
compressible zone of less compacted soils next to the wall that would
radically alter the earth pressure distribution and the mobilization of
soil strength. A general discussion ensued involving Raymond. Bonaparte.
McGown. Jewell and Murray based yet again on the topics of strain compati-
bility between soil and both extensible and inextensible reinforcements
and the need for knowledge of horizontal strains.

Gourc then compared a number of different methods of analysis indicating

the effects of variations between the methods and discussing the topic of
partial factors of safety. Andrawes and McGown pointed out the difficulty
of applying different partial safety fac,tors to the geosynthetic and to
the soil and still expecting to obtain strain compatibility between the
components. Great care would need to be exercised in selecting the
different safety factors.

Rowe presented his contribution which dealt with the general methodology
for deve'loping a successful finite element analysis for reinforced soil
walls. He suggested that finite element analysis was sufficiently complex
that it is unlikely to become a day to day design tool but is likely to be
used in the development of parametric studies and design charts. One
point of discussion of importance to most analyses was the need to
simulate the construction process and thus avoid the "switch on gravity"
approach in which gravitational forces are applied only to an idealized
complete wall model.

In summary during this and most other sessions there was general agreement
that <Pcv was the best and safest strength parameter for general design
usage. Limit equilibrium analyses still provide an acceptable simple
method, for general design and that there is hope to improve them through
better comprehension of ~ the aspects involved. This includes the soil.
the polymer and their interaction. Finite element analysis can be a most
valuable aid in developing parametric studies and making the best use of
the limited amount of case study information available. Formal contribu-
tions to this session were provided by Scott. Floss and Milligan and they
follow this review.
FINITE ELEMENT ANALYSIS OF REINFORCED SOIL

R. Chalaturnyk, D.H.K. Chan, J.D. Scott

Department of Civil Engineering

University of Alberta
Edmonton, Alberta, T6G 2G7

This summary outlines the development of a finite element

program capable of analyzing the performance of reinforced
soil structures. SAFE (Soil Analysis by Finite Elements), a
computer program developedl at the University of Alberta (Chan,
1985), was selected for the development of the program. The
modifications to SAFE include the implementation of a two
dimensional, isoparametric bar element specifically suited to
modelling the behaviour of geosynthetic reinforcing materials
and an interface element for modelling the soil-reinforcement
interaction. SAFE is based on a displacement formulation
assuming small strains and small deformations. The program is
capable of performing two dimensional, plane strain analyses
using total or effective stress formulations for either fully
undrained or drained soil conditions. Table 1 lists some of
the soil, reinforcing and interface models available with
SAfE. Load increment subdivision, program restart capability
at any stage of an analysis, material property variation at
any stage of an analysis and choice of stress calculation
procedures are some of the standard features incorporated
within the program. An element birth and death option allows
incremental construction analyses to be conducted.
Sever"al post-processing programs have been developed in
order to aid in the examination of the finite element analyses
results. finite element mesh and deformed mesh plotting,
s tress and strain contour plotting, displacement arrow
plotting, reinforcement load distribution plotting and
interface normal and shear stress distribution plotting are
all available. Additional development of SAFE includes no
tension analyses using a crack model, an anisotropic
plasticity soil model, special shear band element with a
discontinuous shape function and a time dependent, strain
softening soil model.
SAFE is currently being utilized in examining the
stability of a steep (1:1) reinforced cohesive soil embankment
constructed on a rigid foundation. A total stress, undrained
finite element analysis of this slope is being conducted in
order to investigate the end of construction behaviour of the
reinforcement and the soil. The intent of the research is to
compare the finite element analysis results with limit
557
P. M. Jarrett and A. McGawn (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 557-560.
© 1988 by Kluwer Academic Publishers.
558

equilibrium results in order to determine whether present

limit equilibrium design methods adequately assess the factor
of safety of a reinforced slope.
Figure 1 illustrates the geometry of the reinforced
embankment selected for the analys is. The f ini te element
discretization of the problem, as illustrated in Figure 2,
utilizes 8 node rectangular and 6 node triangular isopara-
metric in soil regions, 3 node isoparametric reinforcement
elements for the geosynthetic material and 6 node isopara-
metric interface elements for the soil-reinforcement
interface. The inset in Figure 2 illustrates a typical
arrangement of these elements.
Line·ar elastic and nonlinear elastic analyses are being
conducted for three cases:
1. Soil elements only (unreinforced embankments)
529 elements, 1,652 nodes
2. Soil and reinforcement elements only (no interface
elements)
797 elements, 1,652 nodes
3. Soil, reinforcement and interface elements
1,333 elements, 2,762 nodes
As a results of the large number of elements and degrees of
freedom required for the analysis, solving the finite element
equations necessitated the use of CDC Cyber 205 supercompu~er
which is located at the University of Calgary.

Figure 3 illustrates the contours of soil strength

mobilized in the embankment at a constructed height of 10 m.
The strength mobilized is defined as the ratio of the
deviatoric stress at failure to the mobilized deviatoric
stress. The shaded zone indicates the region where the
available soil strength has been exceeded. Contours of
different stress and strain parameters can be plotted using
SAFE's 'post-processing programs. Figures 4 and 5 illustrate
the normal and shear stresses along a reinforcement layer,
which is 3 m above the foundation, at embankment heights of 4,
6, 8 and 10 m. Both the top and bottom soil-reinforcement
interface stress distributions are plotted on the graphs. It
is clear from Figures 4 and 5 that the top and bottom
interfaces are behaving identically throughout the embankment
construction. Figure 6 illustrates the development of the
tensile load along the length of the reinforcement during the
embankment construction. The progression of the point of
maximum reinforcement tensile load as the embankment is
constructed is clearly illustrated in this figure.
REFERENCES

1. Chan, D.H.K. 1985. Finite Element Analysis of Strain

Softening Material. Ph.D. Thesis, University of Alberta,
Edmonton, Alberta, 345p.
 559

Table 1 Element Material Models

.SQI Bei10rrerrert ~
( Stress - Strain) (Load - Strain) (Shear stress- Displacement)

1) Lilear EIcSic 1) Lilear Elastic 1) Linear Elastic

2) Hyperbolic Elastic 2) Nonlinear Quadratic 2) Hyperbolic Bastic
3) Elastic Perfectly Plastic 3) Elastic Polyromial with linear failure
or BritIe Plastic; erwekJpe
-vonMises 3) Hyperbole Elastic
-Tresca with wved faikJre
- Dnd<er - Prager envelope
- tAllY - Crubrrb
( associaIed anj ron-
associated flow rule)
4) Elastic Plastic Strain
hardening anj
softening models
5) Elastic Hyperbolic
softening model

RlgldFoundatJon

Figure 1 Reinforcecl Errbankment Geometry

-.
I ,.".,.. RemrI SM,

-- i
~\

-
--~ .
l-tii~·i"I· '
--
AIInfCIrc:ement a.r..nt--+--"

...

Figure 2 Rnite Element Mesh of Reinforced Embankment

 lJo
0\
MOBILIZED STRENGTH CONTOURS (%) 0
Reinforced Embonkment: With Interfoce Elements
REINFORCED EMBANKMENT; ELASTIC; ALL ELEMENTS
LINE 0/0
I 10 • N.4m a*
2 25
i• 'iI",.,;;",;;;
3 50 .-';;'-i;;; ~
4 60 'i"H";"iij';
~\ ~cf
5 70
6 80 '"
g:f
7 90
100
r - -- - ----- ----. ~~
;;,'-;c---->'
.
8 '! ) -..-'-1
3 .:..~'-----""';, ~ all
c 9 125
.. g ~

.! 55
\1 < zseel '--'-'-'---'.:.~ I!~
c \ '. """
o
~\~:.................:::::::...:.:::::.:::,:.:.....:::::,:::..... .... ,
II
.c I
e ~
w50 --.. lio 120
• I
, i i i
: :,;, TO
i!l

100 X-Coordinate. m 2' 20 15 to 0

Distance from Foce of Embankment. m

Figure 3 Strength Mobilized for Embankment Height = 10m

Figure 4 Irterface Normal Stress Distribution

REINFORCEO EMBANKMENT; ELASTIC; ALL ELEMENTS

REINFORCED EMBANKMENT; ELASTIC; AU ELEMENTS ~

~
• M_4""
"'H'~'.'~'" • E
'ii"H'~'ii';;; ~ -H-:'-';'--
ii·H;·i~'­ i
H .. IO,..
'i"'H~rom i
:;~ / ' ------- ' ~ E
- +... g:f
:l~
Vl
~ 3m '
/' /
//,\" '\
3 m ....... 0.0"0111"'- ~ V / ...... \\.\
./",-:::::= ~1 , /'
J" .5
~'" ,/ '
.. ... \'\\\". '8
-'_/ ,p:-::!i. -/ -'/ ...... , ... -.3
g - ~ \, .
,~:::::::::---- ,/,.;'
."..- -'- -- - ......... ,"'S-'
-::::::::::"~""""""~,::,"''', ,," ' \\';;1 a

i;;'
-----
.. ,.:3:::-::;:-::=:7:::::::;:::::;:=·····..·····
t5 IV
!8
o
a \ 25 20 '5 10
Ohdonce from Foce of Embankment. m
0

" Distance from Foce of Embankment. m

Figure 5 Interface Shear Stress DistrbJtion Figure 6 Load Development in Reinforcement

REINFORCING ELEMENTS IN STEEP SLOPES AND VERTICAL-FACED EARTH
STRUCTURES - German State of the Art
R. FLOSS
Technical University Munich

1. GENERAL DEFORMATION AND FAILURE MECHANISM

The following basic variants are considered:
(1) Earth retaining structures with wall elements as face
lining and structurally connected reinforcement similar to the
"Terre Armee" construction method
(2) Same as (1), with gabion facing
(3) Fill with reinforcing mats (fabric, vleece with tensile
strength), which are folded back at the face so that the soil
cannot slip out; constructed (a) as so-called bolster dam
(Fig. 1) or (b) as safety measure for extremely steep slopes.

system of equations for limit state

(1) soil: <5, =<5z = o'z
-l_l-sin.p
<53 =G"x + L!.<3"x = ~'f '<3", A'f- 1+ sin <p
(Z) reinforcement:L!.<3"x = Z· Z(xi ·L!.x/o
0= Z·zu[ Z· L!.x/L!.G"x
(3) soil - reinforcement - interaction:
Z(xi = ZU[ Z = <5 z ' ton 'PR
FIGURE 1 Interactive forces for the limit state of a steep
fill slope constructed according to the principle of the
so-called bolster dam

561

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 561-567.
© 1988 by Kluwer Academic Publishers.
562

As a rule, non-cohesive soils which are not prone to defor-

mati on by creep and can e a si 1y be d r a i ned are used for the
construction of earth retaining structures. For fills, how-
ever, one can also use cohesive material.

For want of a reliable empirical and measuring data per-

taining to the failure and deformation behaviour of self-
supporting earth bodies, whose high bearing capacity is
achieved by soil reinforcement interaction, a number of ques-
tions remain unanswered. Besides, the applied construction
methods may differ substantially and other conditions may have
some influence, too. Previous observations and measurements of
embankments constructed as bolster dam indicate no dispro-
portionate displacements, creep or other damages to fabric or
vleece in the face-area, providing there is a high degree of
compaction of the non-cohesive soil. The bearing capacity is
partially much higher than the conventionally determined ulti-
mate loads. In view of these findings, it can be taken for
~anted that the plane-line reinforcing effect and the full
plane activated shear stresses in the non-cohesive soil can
produce a somewhat more effective reduction of the earth
pressure in the case of the bolster dam than in cases where
the system is reinforced with bands.

According to general soil-mechanical knowledge, such compo-

site soil systems act as a slack body with internal shear,
with the reinforcement layers creating the effect of an aniso-
tropic cohesion. There is a certain analogy to the principle
of the cofferdam which also acts as a slack body, where the
external forces are transmitted to the base by means of shear
forces between face and fill.

Compared to the monolithic, rigid body the conventionally

reinforced composite body is noted for its totally different
settlement, earth pressure and failure behaviour (Fig. 2):

(a) Because of its relatively low natural stiffness the

reinfo'rced composite body reacts less sensitive to differen-
tial settlement and horizontal deformation. From the cross-
section it can be noted that the settlements occurring under
the composite body are trough-shaped, i.e. the maximum settle-
ments occur under the body rather than under the face-line if
uniform subsoil conditions prevail.
(b) The effective, oblique eccentric force acting in the
base and resulting from the dead weight of the composite body
and the earth pressure, causes a bearing pressure a O which
compared to the monolithic rigid body - lessens towards the
outside.
(c) Contrary to the monolithic body, the earth pressure
does - not act at the backside. Instead, the forces resulting
from earth pressure are reduced within the composite body,
through shear stresses and arching effects in between the
reinforcing elements so that only a residual pressure reaches
the face area. This was confirmed by measuring results from
 563

tests on supporting systems of the "Terre Arm~e" construction

method. The borderline of the zone of active earth pressure
follows the geometry of the maximum tensile force of the re-
inforcement; this is a curved line which embraces a much
smaller zone than that of the plane sliding body according to
the earth pressure theory by Coulomb.
(d) The most important factors of the stability analysis
are the safety against base failure, sliding of the composite
body, the body cannot tilt but would come apart in the case of
an effective moment of tilt. Mathematical analyses show that
the composite body by itself (without vertical load) has suf-
ficient inner safety against failure if the stress redistri-
bution caused by tensile forces of the reinforcement is con-
sidered and specific geometric design criteria are observed.
Under these conditions, and presuming that the safety against
base failure is observed, it would suffice to investigate 'only
potential shear zones outside the reinforced body to verify
the safety against slope failure. However, if b is much larger
than h the slack behaviour of such a broad body would also
1e ad t 0 i n t ern a 1 s tat e s 0 f f ail u r e wh i c h s tar t 1 0 cally. Su c h
cases require special investigation.
As mentioned earlier, the known investigations show that the
bearing capacity of the composite body can be well above the
conventionally calculated failure load. The following pheno-
mena may add to this behaviour:
1. The reinforcing layers cause global and local prestress-
ing effects in the non-cohesive soil.
2. The construction of any new reinforcing element changes
the resultant main stress direction and thus the geometry of
the potential failure figure.
3. Because of the restraint between reinforcing layers the
soil is compacted to a higher degree.
4. The horizontal load transmission by reinforcing layers
and the compaction effects lead to vault-type force transmis-
sion links in the non-cohesive soil and zones of high shear
strength along the reinforcing layers; thus additional hori-
zontal stresses ax induce retaining shear forces.
The influence of the cited effects is likely to be parti-
cularly high in cases of geotextile reinforcing elements be-
cause of their planelike friction and bonding effects.
Be sid e s, the rei s noway 0 f r eli e f for the so i 1, be c a use 0 f
the all~round restraint. The vaUlt-type load transmission
structure within the soil strengthens layer by layer as the
construction processes. Within certain areas they even relieve
the reinforcement. In areas where the limit state is reached
locally, these load transmission links are changed as a result
of the ,newly induced a -stresses as parts of the load are
transmitted to the rein>forcement, so that the plasticizing
process is delayed. However, this process implies that the
az-stresses are not excessive.
564

The above-mentioned phenomena need to be verified by future

in-situ investigations and models. They are still mere hypo-
theses, but observations indicate that they are close to real-
ity. Considering the relevant technical and safety demands,
there is little doubt that it will be possible to construct a
composite body of uniform and high density.

(1) sl ack body (Z) rigid body

r--B~
x x
I <DG
J...
L>E
IT

E, Ez
'15'0
Smox Smox

, --- Go

z
S' ---+- Ze -+
<D active zone Z
B' =B-Z·e E,<Ez(tono.OJ ~o = Yt. 'l'iz
+

6 0 = V/ S' (J) passive zone Zj = L!. Ez

Zj = L!. E, x, > Xz
FIGURE 2 Comparison of models: (1) slack body and (2) quasi-
monolithic earth body with regard to their settlements, base
pressure a O' earth pressure E, and tensile force of reinforce-
ment Zi'

2. CONSERVATIVE ANALYTICAL MODEL

Safety aspects and the lack of reliable empirical data are
the rea.sons why the effects descri bed inSect ion 1 have not
yet been reflected in the stability analysis. As a substitute
one resorts to analytical models which are based on a quasi-
monolithic composite body (Fig. 2). For the typical case,
these models abide by the following design principles, which
are on the safe side: (a) application of the earth pressure,
according to Coulomb, to the backside of the composite body,
but (in contradiction to the monolithic precondition) with an
evenly limited active wedge of failure under ,J- = n/4 + <1>/2
within the composite body; (b) induction of the fetaining ten-
sile forces of the reinforcement behind the shear zone; (c)
trapezoidally distributed base pressure analogously to the
rigid body which is subjected to effective, oblique eccentric
load; . (d) sustenance of the horizontal earth pressure compo-
nents in the base of the composite body; (e) in the case of
face elements, design of the elements with full earth pressure
application.
 565

The conventional stability analysis includes proof of the

external and internal stability of the composite body.
(1) Depending on whether a steep slope or supporting earth
structure is being considered, the safety of the external sta-
bility has to be proven by:
(a) Safety against sliding in the base of the composite
body according to DIN 1054 and for supporting structures with
stiff reinforcing elements (e.g. geo-grids), also in each re-
inforcement joint. Safety coefficient n > 1.2 1.5 depend-
ing on the loading case. g
(b) Limited eccentricity of the load inb the foundation
joint.
(c) Safety against base failure according to DIN 4017 .
Safety coefficient n > 1.3 ... 2.0.
(d) Safety againh slope failure according to DIN 4084;
safety coefficient n > 1.1 ... 1.4. Depending on loading case
and analytical method, also for parts of the reinforced earth
body (especially for systems with vertical and traffic loads).
The investigations must also include the construction condi-
tions. Although check-ups with a lower safety coefficient
still yield a statically sufficient safety, the possibility of
conspicuous deformation occurring in the system cannot be
ruled out. Shear and slope failure verified according to DIN
4084 only apply to plane deformation states; applied to spa-
tial cases they yield results which are on the safe side. Dur-
ing construction of the composite body higher earth pressures
may occur as a result of the layer-by-layer compaction of the
soil; they have to be considered in the calculation.
(2) The safety of the inner stability has to be proven by:
(a) Safety against tensile failure of the reinforcement
(n > permissible ZIZ.).
(b) Safety again1t pUll-out of each individual reinforce-
ment layer, with the retaining forces over the embedded length
of the reinforcement behind the sliding wedge and with a re-
duced coefficient of friction for the interactive forces bet-
ween reinfvrcement and soil.
(c) Safety against sliding in horizontal reinforcement
levels within the composite block with reduced coefficients of
friction according to soil and reinforcement type.
(d) Sustenance of tensile forces at transition points bet-
ween rei nforcement and outer 1 i ni ng (Z.1 > 0.85 max Z.).
1

In the working state it is essential to define the permiss-

ible tensile force (permissible Z) for the respective rein-
forcement. This working-tensile-force depends upon the service
life and elasticity of the reinforcement and its long-term
behaviour (ageing, temperature-dependent reductions etc.).
Although .the experience with geotextile reinforcement is still
rather limited, the following guide values may be used in
practice: They depend on the planned service life of the
structure:
(a) Permanent reinforcement: sustaining tensile forces is
20 % to 25 % of the tensile force at 5 % permissible strain;
566

(b) Temporary reinforcement: sustaining tensile force is 30

% of tensile force at 10 % permissible strain.
3. MODIFIED STABILITY MODELS
Although there is a number of modified models, it is only
possible to summarize their principles in this paper. Contrary
to the conventional approach, however, they are all aimed at
describing a bearing condition which compares more favourably
to the natural state.
(1) Modification of the geometry of the active failure
zone: (a) assuming a broken limiting line for the active slid-
ing body for better adaption to the true course. The most un-
favourable shear zone is found by variation of both angles of
sliding surface. (b) Assuming a sliding body geometry, which
approximately follows the geometry of the maximum tensile
forces (e.g. Berg 1986, John 1986). (c) Using a block element
model with application of tensile force at the edges of the
elements (Fig. 3). Not only does this method consider the tan-
gential component of the tensile force T, as retaining force
but - with component T2 - it also considers the retaining
shear forces resulting from the normal stress portion which is
increased by reinforcement in the failure zones.
(2) Assuming reduced earth pressure in face area, giving
consideration to the reduction of earth forces within the re-
inforced slack earth body.

!llllill

forces in shear zone

Tz=N· ton y;

\: reinforcement

FIGURE 3 Block model for steep slope with two elements, and
showing approach of tensile force distribution in the failure
zone
 567

(3) Application of higher shear parameters <p and c' for

the determination of the bearing capacity of the composite
structure (e.g. Ingold/Miller 1982).
(4) Modification of failure mechanisms based on the des-
cribed vault-theory. Instead of assuming a uniformly curved
sliding zone for the entire composite body it is implied that
the system fails layer by layer (Werner/Resl 1986).
REFERENCES

1. BERG, R.R., BONAPARTE, R. et a1 (1986): Design, Construc-

tion and Performance of Two Geogrid Reinforced Soil Retaining
Wall s. 3 r dIn t . Con f. 0 n Ge 0 t ext i 1 e s, Vie n n a, Vol. I I, 4 0 1 -4 0 6
2. IN G0 LD, T. 5 ., MIL LER, K. S. (1 982 ) : An a 1y tic a 1 an d La bo r a-
tory Investigations of Reinforced Clay. 2nd Int. Conf. on Geo-
textiles, Las Vegas, Vol. III, 587-592

3. JOHN, N.W.M. (1986): Geotexti1e Reinforced Soil Walls in a

Tidal Environment. 3rd Int. Conf. on Geotexti1es, Vienna, Vol.
II,331-336
4. WERNER, G., RESL, S. (1986): 5tabi1itatsmechanismen in
geotexti1verstarkten ErdstUtzkonstruktionen. 3rd Int. Conf. on
Geotexti1es, Vienna, Vol. II, 465-469
ANALYTICAL TECHNIQUES AND DESIGN METHODS - DISCUSSION

DR. G.W.E. MILLIGAN,

University of Oxford, England.

In considering the possible development of horizontal strain in a

reinforced soil wall, it should be noted that in certain cases the soil
strains may be simply related to the movement of the wall facing.
Bransby and Milligan (1975) and Milligan (1983) have shown that the
strains behind a retaining wall are related to the movements of the wall
as shown in Figures 1 and 2. Comparisons with the results of model and
full scale tests have confirmed that the analysis works well for outward
translation and rotation of the wall about its base. For outward
bulging, or rotation of the wall about its top, the simple analysis
breaks down and the strain field becomes more complex.

A reinforced soil wall with a full height facing normally deforms

by rotation of the facing about its base. The maximum strains likely to
be developed in the soil may then be given approximately by the simple
analysis. The presence of reinforcement within the soil will modify the
strain pattern somewhat and reduce the maximum strains.

Two examples are considered here: -

(i) In the full height panel wall at RMC, the

outward rotation of the facing panel at the end
of the test was about 4.3 x 10 - 3 radians. The
peak angle of dilation for the fill was about
20° or a little less (the calculation is rather
' insensitive to the value of ..p). The resulting
tensile horizontal strain at all points in the
active zone is then calculated to be about 0.6%.
This compares with measured tensile strains in
the reinforcement in the active zone of about
0.4% to 0.6%.

(ii) For a more typical full scale wall, the angle of

dilation for the fill might be 10°. If the
allowable outward rotation of the wall facing is
1%, which would normally be considered a fairly
large movement, the horizontal strain in the
active zone is 1.2%. The reinforcement strain
cannot normally exceed this value, and this
would then provide a strain limit criterion for
the wall. If it is desired to allow the
reinforcement to develop larger strains, either

569

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 569-570.
© 1988 by Kluwer Academic Publishers.
570

larger facing movements must be tolerated or

additional soil strains must be induced by the
presence of a compressive layer behind the
facing or other similar construction expedient.

References

Bransby, P.L. and Milligan, G.W.E. (1975). Soil deformation near

cantilever sheetpile walls. Geotechnique 25. N°2.

Milligan, G.W.E. (1983). Soil deformations near anchored sheet

pile walls. Geotechnique 33. N°l.

Strains in hatched zone

i' = ¥
shear strain = Z() Sec.
volumetric strain = -.l.() tiu-.. r= v

'f' = angle of dilation of soil

f) == angle of rotation of element of wall

.compressive strain positive

F;g.1 Soilstrains behind retaining wall

----______~------~~--_r~~--~c.
horizontal strain -/E"

Fig.2 Mohr's circle for strain increment

Construction Methods and Economics
 CONSTRUCTION METHODS. ECONOMICS AND SPECIFICATIONS

C.J.F.P. Jones, Professor, University of Newcastle upon Tyne

A. McGown, Professor, University of Strath.clyde
D.J. Varney, Engineer, Netlon Ltd.

1. INTRODUCTION
The modern forms of reinforced soil have evolved from an
understanding and introduction of effective .construction techniques,
the first of which was developed by Vidal (1966). The first
structures used steel strip reinforcements arid it was not until the
early 1970' s that polymeric reinforcements were used successfully
with vertical faced structures. Prior to this, high density
polythene grids had been used in the construction of railway
embankments in Japan, Yamamoto (1966).
This paper considers the practical aspects of the construction of
vertical or near vertical reinforced soil structures formed using
polymeric reinforcements. Included are details relating to
construction systems and techniques, economics and specifications.
Although there are numerous other possible construction arrangements,
those detailed here have all been used in practice. Where possible
the potential advantages displayed by individual systems or materials
are described.

2. CONSTRUCTION METHODS
Reinforced soil structures must be of a form in keeping with the
assumed idealization and analysis, however, the theoretical form of
the structure may be quite different from the economical prototype,
and attention should be paid to the method of construction throughout
the design process.
Speed of construction is usually essential to achieve economy and
in part this may be achieved by the simplicity of the construction
technique. Construction techniques compatible with the use of soil
as a constructional material are required. The use of soil,
deposited in layers to form the structure, results in deformations
within the soil mass caused by gravitational and compaction forces.
These deformations result in the reinforcing elements positioned on
discrete planes moving together and being tensioned as the layers of
soil separating the planes of reinforcement are compressed vertically
and expand laterally. Construction techniques capable of
accommodating this internal consolidation and straining of the fill
are required.
Failure to accommodate the compression, particularly at the face,
may result in loss of serviceability or even rupture of
reinforcements or connections, whilst restriction of lateral
expansion of the fill may prevent the tension in the reinforcements
developing fully.

573

P. M. Jarrett and A. McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining StructlUes, 573-611.
© 1988 by Kluwer Academic Publishers.
574

2.1 BASIC TYPES

Three constructional techniques which can accommodate vertical
settlements and limited lateral movements within the soil mass are
shown in Fig 1. Except for some special cases, reinforced soil
structures constructed above ground use one or other of these forms
of construction or a combination of them.
2.1.1. Concertina Method. The constructional arrangement of the
concertina method developed by Vidal (1966), is shown in Fig lea).
Differential settlement and lateral movement wi thin the soil mass
(dl-d4) is achieved by the front or face of the structure
concertinaring in a manner similar to the action of a set of bellows.
Some of the largest modern reinforced soil structures have been built
using this approach, and it is the form of construction frequently
used with geotextiles and related reinforcing materials in both
embankments and cuttings. A flexible hoop shaped steel facing unit
has been ·used when the structure is reinforced with strip
reinforcement, Fig 2(a), as have facing units formed from aluminium
and glass reinforced plastic, although the latter have had limited
success due to their lack of robustness.
Geotextiles and related materials often provide their own facing
by employing wrap-around techniques, Fig 2(b),(c),(d) and (e). This
is particularly useful for structures where distortion of the facing
is acceptable.
2.1.2 Telescope Method. In the telescope method of construction
developed by Vidal (1978), Fig l(b), the deformations within the soil
mass are accommodated by the facing panels closing up and moving
forward an amount equivalent to the internal deformations. This is
made possible by supporting the facing panels by the reinforcing
elements and leaving a discrete horizontal gap between each facing
panel, i. e. the facing panels hang from the reinforcing elements.
The horizontal gap between each facing panel may be effected by the
use of compressible gaskets, Fig 3. Failure to provide a large
enough gap between facing elements can result in crushing and
spalling of the units as the fill deforms under the action of gravity
and compaction forces.
The closure between panels will vary from application to
apRlication depending upon the geometry of the structure, quality of
fill material, size of the facing panels and the degree of compaction
achieved during construction as well as the subsoil conditions.
Typical movements on a steel reinforced structure, reported by Finley
(1978), show vertical closures of 5-15 mm for facing panels 1.5 m
high. Also the shape and form of the facing panel must be compatible
with the procedure adopted, and reinforced concrete cruciform, tee-
shaped, hexagonal or "Z" shaped panels covering 1-4 m2 and 150-250 mm
thick are typical. The construction sequence for the telescope
method is shown in Fig 4.
2.1.3 Sliding Method. In the sliding method of construction
developed by Jones (1978), differential settlement and compaction
within the fill forming a reinforced soil structure can be
accommodated by permitting the reinforcing members to slide
vertically relative to the facing. Slideable attachments can be
provided by the use of grooves, slots, vertical poles, lugs or bolts.
If vertical poles are used, these may form the structural elements of
the facing and the facing may become non-structural providing only a
 575

(c) sliding method

Figure 1 Methods used for constructing reinforced soil

structures

(b)

(e)

101

I.)

(a) R(infon:cd $Oil wilh uniu

(b) Method of constNC1ion. fabric wall
(c) Method or construClion. r.bric wall
(d) M(thod of construction of sloping traven.c
(e) M(lhod of construction. fabric or gco,rid or
cutting

Figure 2 Concertina Method of Construction

576

Facing unit
Temporary wedges
Reinforcement

Figure 3 Telescope Method of Construction

.Iev.tion concrete
foating
Con'll'lIction sequence - Step 1
Step I: Casl rOOlinl and dninaae ,approximately
I SO JC 300 mm) with top surface lnocl. Erect half' panels. Erect
fuU s.ize panek and rll temporary wed,es to creale horizontal
pp between units. Clamp .djacmt units IOJCthcr and prop
fim rows from front.

clamps to mlintlin
.lign~ent =~f :~~!
I

·IT--~'~__L~______~_.
1Omoo"'" _ \

elev.tion

Construction scqutnee - Step 2

Step 2: PIa« fill and compact to k\lcl of first fO ..... of
mnron:tTDenl A·Acompaction within 2 m ofthe 'ace must be
undertaken with care so as nOI to crClte extrssivc distonion
of the r.cin,_ Place reinforcement and ,tllch (0 r.cinl. nole
any form or tC'inroR:cmcnl may be used.

Construction sequcnc:r - Sicr )

Slep) ""'hen ~lJlin, has rc.achrd IneIIA-AI ~mo ... c clamps PLaa IDolher row of panels and wed,ts RcpJ,aa c1.1.mp~ I' h&Jhcr
lc\d and conllnue cycle. Ai erection prOCC'Cds. femo\"e lemporat) a'Cdcn 10 pennll vcnal Klllcmcni or JOII mau and racing
."",,., When p.lensiblt fClnronxmcnl IS Us.ed .• de,rft Dr prcstrnsm, Is. rcqulrt'd pnor 10 pbciDJ fill
Figure 4 Construction seq~ence for the Telescope Method
 577

covering, whose role is to protect the completed structure from the

elements, prevent erosion or act as an aesthetic feature. With this
arrangement the type and form of the non-structural part of the
facing can be chosen to suit the particular application or
environment, Fig 5. The erection sequence for a non-structural
facing is shown in Fig 6.
If a structural facing is used, the connecting element, the
vertical pole, may be reduced in size to an appropriate form, such as
a bolt; alternatively, attachment of the reinforcement to the facing
can be through slotted holes. The facing may be rigid or semi rigid.
Up to a height of approximately 10 m a rigid facing may be used; for
heights above 10 m an elemental form of facing is appropriate. Where
a full height rigid facing is used, this is erected and held in place
before filling starts. The erection sequence is shown in Fig 7.
2.1. 4. Hybrid Methods. Recently hybrid structures, using :a
combination of the Telescope and Sliding methods of construction have
been used successfully with geogrid reinforcement, Fig 8. In this
system the reinforcement, which passes through the facing, is
attached by means of a slideable connection to king posts, whilst the
facing operates in a telescope arrangement. The primary advantage is
of economy and speed of construction. The system has been used to
provide masonry faced structures of high aesthetic appeal. When
architectural features are omitted this form of structure is
particularly suited to industrial situations. In the case of unfaced
structures the reinforcement is connected to the rear flanges of the
king posts so as to avoid the risk of vandalism and degredation of
the reinforcement.
2.1.5 Specials. In some special conditions there is no need to
make allowances for consolidation of the fill, as the degree of
movement involved is limited by the design. This can occur when the
foundation beneath the structure is very competent (eg. rock) and
when the backfill material used is of very high quality and well
compacted. In these special circumstances the differential movements
(dl-d4) in Fig 1, are very small and may be safely accommodated
within the slackness of the connections and the extensibility
inherent in polymeric reinforcement. The use of stiff reinforcement
and rigid ' connections is not recommended with this form of
construction.

2.2 SYSTEMS IN USE

2.2.1 Reinforced Fill Behind Conventional Walls. The benefit of
reinforcing the backfill of conventional structures has been
demonstrated by Saran et al (1979). In this technique the active
pressure on the structure is reduced. However, the structure is
reliant upon the facing to prevent erosion and for stability of the
fill immediately adjacent to the wall.
2.2.2 Independently Reinforced Soil Mass and Facing Panels. An
advance on the use of a reinforced backfill with conventional
structures can be achieved by substituting an incremental panel to
act as the facing. In a recent structure built at Lithonia, Georgia,
small precast concrete cruxiform units into which were embedded two
102m long geogrid "tabs" were used as the facing. The "tabs" were
used to secure the facings to the independently reinforced soil mass,
and the main geogrid elements were simply laid horizontally in the
fill without any connections to the facings.
578

vertical poles

reinforcement

Figure 5 Sliding Method of Construction

nep 1

SKlion
1.' elellJlion
, concrete footing

Sr,p J
Cast Coolin! approximately nODx 1SO) mm 'With lOP surface Intel ErClCt ~cnK:.al reinforcement pOles.
[,oct halt uniL Platt porous dr.lina,c pipe.
Erect first full units. PUtt no fines concrele.

step 2
Ibl
section A-A stchon 8-8

.dd '"'' of fill q~$$$77$5m5W//'

~,~,m .. et '.ye, of ¢w%«w/;
ac!d unit and
on s.tetioll A-A.

COmpaC1101ytr of
hit 'el

.~d f-'Ing unit and

'Klion 8-8

SIt/, 1
Note: Rcin(orcancnl poIittoD is al mid·hei,pn of facin, umt
Any fonn or rclnJ'Ota1Dml mlJ be lUCId,
Com~ctioD or the rill dost 10 the rlee is restricted 10 s.m.all plan. so u not 10 dillon the- ficin,
Spt'td of c:onSlrl,IC'tMJn is dC'pC"ndnsl upoa the speed or platLnJ Ibe fill
A bokf raein, is uwally used to dis,utle any incolULutorio Of dsslonlons c.au~ durin, conllNC1ion
Construction can be Itopped.1 .n,.lnrl or position .. ilbout .ny fcar (or the wet,. or Ihe \!Work en

Figure 6 c:n"lt"~~rt~n Sequence of Sliding Method

(non-structural facing)
 579

b~
~T"~
~~- -;
1/:--
I ~
\
~ m'n50mm

galvanised
!tHltublng
nut

Detail 'A'

Slag, J:
Cast fooling (appro. 1SO x 300 mm) plus upsland. Elocl and
prop facing panels
concrete footln9
/ Slagt 2:
Construct drainage.
Place fill layers and compacL
When fin level with top of first pair of connecting lugs attach
first level of reinforcement.
It!'mpor.lIy folding wwgt!' Continue filling.
When filling is complete, or when sufficient fill has be.n
fbi stage 1 placed to stabilise facing lemove props and folding ... edges.

------'\
------1
de-tail 'A'

lei SlClge '2

Sliding molhod of construction - rigid facing

,

Figure 7 Sliding Method of Construction - rigid facing

 580

(jf
limit of rein-
M24 bolt forced soil

lass fibre rein-

forcing strap
sponge to allow
downward movement ,.-~
..' :......_'re'n'torced concrete ' of strap with :,.-
string course at core / tie beamENlAAGEMENT OF settlement of fit!
footpath level of STAAP FIXIN G .... --; .....
. .-;;;: ..... .

masonry facework

~
mass concrete
baCking to
masonry

reinforced-concrete
footing
. - ~para~:e~etween
"'~"",,,,--- fiilll and drainage layer
reinforcement laid on
j These layers
only
necessary if
PFA is used

Figure 8 Hybrid Method of Construction - Telescope and Sliding

 581

This structure has a maximum height of 6.1 metres and forms part of a
material handling platform supporting 450kN dumptrucks. A cross
section through the wall is shown in Fig.9. The design and
construction of the wall has been described by Berg et al (1986).
2.2.3 Incremental Construction with Discrete Panels.
The most common form of reinforced soil structure is the telescope
method using incremental panels, Fig. 3. A number of proprietary
systems are based upon this technique, including the familiar
Reinforced Earth Company's system which uses a reinforced concrete
cruciform facing unit and the Georgia Department of Transportation
GASE method, which uses steel grid reinforcement. The method has
been used with polymeric reinforcement in the United Kingdom, the
Middle East, and recently in Japan, using Tee shaped and 'Z' shaped
concrete facing panels.
The Soil Structures International Limited 'Websol' Fig.4 A number
of the most sucessful proprietary methods which uses polymeric
reinforcement. This is based on a precast concrete facing and tape
reinforcement, Paraweb, formed from ten lanes of high tenacity
polyester fibres encased in a sheet of low density polythene. The
Paraweb reinforcement, which is connected to the facing panels by
toggles, is laid using an entire roll at a time, typically 100m.
long. This is acheived by lacing the reinforcement between the
facing and a restraining member laid in the fill, Figure 10. In some
structures layers of , Terram' geotextile are laid on top of the
reinforcement, Kempton et al (1985). Recently, an anchored earth
system using paraweb has been introduced in Austria. This system is
now being used in the United Kingdom, Lazlo et al (1995).
Geogrid reinforcement has also been used with the incremental
construction system. Pigg and McCafferty (1985) have described an
application in the UK, whilst Berg et al (1986) have detailed the use
of geogrid reinforcement with the Georgia GASE system. Recently
geogrid has been used with the Japanese 'z' panel, Yamanouchi (1986).
2.2.4 Full Height Panels. The use of full height panels is often
an attractive form of construction for walls of limited height «10
metres). However, since the possibility of articulation by either
the concertina or telescopic methods is lost, consideration must be
given to p~oviding sliding connections, Figure 7.
In the United Kingdom a number of agencies have successfully used
full height panels for highway and industrial projects.
Consolidation of the reinforced fill has been accommodated by use of
sliding connections between the reinforcement and facing panel.
Three polymeric reinforcing materials have been used namely,
"Fibretain" glass reinforced plastic stips, "Tensar" geogrids and
"Paraweb" tape. The system of full height panels was also used for
the construction of the first Anchored Earth structure constructed on
the Otley by Pass in Northern England (Jones et aI, 1985).
In the United States some full height panel wallsconstructed with
rigidly connected steel strip reinforcement have exhibited severe and
unacceptable distortion of the facings on account of there being no
provision for relative movement between the reinforcement and panels
as the fill settles and deforms laterally. Crushing of steel facing
panels has also been observed in Japan on Anchored Earth structures
when no provision has been made for movement of the reinforcing
anchors relative to the facing. Observations on full height
582

U ~~"'M' "
r~ ~~""'r. ~;.~~:::'.. c

~; ~EOGRID REINFORCEMENT rr~PI

L - 3.6m
I

1.3m

:1.r REINFORCED ALL ~1 ,~ ,

':~:l RETAINED ALL: '.':

, MIN. 75mm SEPARATION

~ ct.' ________ . __ _ ~l~~~

~·,,:;w.~,"'S·r:·~= LEValNG PAD " .,.••' •• :
~ """V.......,,~
Lithonia Cross Section
(wa 11 sect i on wi th independent fac i ng)

Figure 9

-
.
I 11-
II
I
I I I , I
[
II
,
ij
, ,. , ,3~ 11
II

JlJ ~
667661

Figure 10
 583

panel structures in the United Kingdom have shown that the

reinforcement regularly settles 20-S0mm with a sliding type
construction (Jones et al 1987). Analysis of this condition shows
that as resistance to this movement develops, so significant stresses
will be imposed on the facing panels.
One of the common construction techniques using geogrids consists
of full height facing panels into which short "tails" of
reinforcement are cast. The "tails" are later joined to the main
reinforcement as shown in Fig. 11. Whilst the relatively low bending
stiffness of geogrids offers more flexibility to accommodate
settlement than does steel strip reinforcement, the system requires
the use of good quality fill which can be well compacted to reduce
settlement and ensure good performance in service. The system has
been used successfully on projects in the United Kingdom and in the
United States at Tuscon, Berg et al (1986).
The 'hybrid' system currently used in the United Kingdom is a form
of the full height system which can be used to provide structures of
high aesthetic quality, Figure 8. As described in paragraph 2.1 (iv)
the system is a combination of the telescope and sliding methods of
construction. Structures using glass fibre, polymeric tape and
polymeric grid have been constructed, Jones et al (1987).
2.2.S Tronderblock System. Tronderblock precast concrete facing
blocks have been used in Norway for the construction of low height
gravity retaining walls for some years, Figure 12. The blocks are
SOOmm long x 2S0mm high x 600mm deep and require a small crane to
lift them into place. Once placed they are stable and self
supporting and provide a suitable facing formwork against which fill
can be placed and compacted. Recently the versatility of the system
has been improved by introducing horizontal layers of "Tensar"
geogrid reinforcement between the rows of blocks which are connected
to the facings. The result is a hybrid structure, part gravity wall,
part reinforced soil, details of which are shown in Figure 13. The
use of reinforcement extends the height to which the system can be
used and structures of S.Sm in height have been completed at Karmay
in Norway. A current development is the introduction of smaller
facing blocks which can be handled without the need for cranage.
A separate development in the United Kingdom is the Porcupine
retaining retaining wall block in which walls exceeding 3m in height
are reinforced with the geotextile "Stabilenka". In poor ground
conditions the use of a 'Nicobag' formation is advised.
2.2.6 Wrap-around Facings. Many permanent structures have been
built which utilise geotextile sheets, grids, nets and mesh
reinforcements as the facings. These may be termed wrap-around walls
which conform to the concertina method of construction. Wrap-around
walls are constructed simply by folding an extended reinforcement
element through 180 degrees to form the face and anchoring it back
into the fill or to another element at a higher elevation. Fill is
usually placed and compacted against external, temporary formwork,
Figs. 14 ~nd IS. The reinforcing elements of wrap-around structures
are usually pretensioned to limit post construction distortions.
Wrap-around facings are a very versatile method of construction and
accordingly a wide range of temporary facing support methods have
emerged, particularly on the initiative of Contractors, who stand to
gain from reduced costs derived from the use of innovative site
584

- G,.~;!.{dtail
/ Public footpath

I
POlystyrene
MHWST
--.lL
Membrane

1/ MHWNT
--lL
40

Facing panel
Plan detail
of facing
panel and joints

Geogrids

Figure 11 Full Height Panel Construction

Figure 12
 585

, =

Figure 13
586

Tensar SR2 geogrid fascia

board

Figure 14

FINE
MESH

REINFORCEKENT

SCAFfOLD FPJ,.....,E
AND BOARDS I!I~--==~f----

Figure 15
REINFORCEMENT

VERTICAL TUBE
LEf"T IN PLACE

Figure 16
 587

specific techniques. Lightweight formwork, comprJ.sJ.ng a grid of

scaffold tubes and boards has been used for the construction of steep
and vertical walls. Fig. 16 shows the arrangement used for the
temporary works of a missile barrier in the United Kingdom.
Sacrificial tubes have also been used to support formwork but these
should be avoided in favour of external bracing, as this reduces cost
and eliminates the difficulties of filling and compaction around
obstructions.
Climbing or sliding formwork has been successfully used both on
steep and on stepped faces, Figs. 17 and 18. These methods assume
the structure will be self supporting as each stage is completed.
In order to avoid surface erosion the facing must be capable of
fill retention once the support is removed. Where open grids, nets
or meshes are used, a turf or seeded mat system with a topsoil lining
is recommended, Fig. 19. This provides a root system which binds the
surface of the fill together; after a relatively short period of
time, the polymeric facing becomes hidden by a mat of grass which
offers protection against UV degredation and damage by fire. Where
aesthetics do not command priority (eg. temporary structures)
retention of fill may be achieved using a geotextile, natural fibre
liner, or coarser fill.
Where geotextile wrap-around structures are specified, account has
to be taken of the facial distortion which may develop. Most
permanent structures are constructed with an outer protective facing,
and the most common solution recommended by geotextile manufacturers
is to provide a 40-75mm thick sprayed concrete facing.
This solution also overcomes the problems of vandalism and
durability, but is not particularly aesthetic and may be expensive.
An example of this method of construction is shown in Fig. 20 which
is a cross section of a 5.2m high geotextile reinforced wall
constructed by the New York Department of Transport. The structure
was finished by applying a 40mm thick sprayed concrete external
facing, Douglas (1981).
Whilst wrap around walls are suitable for permanent works, their
use as a construction expedient is equally common, particularly on
account of the high rates of construction which have been achieved.
Construction times have been reduced by between 30 and 60 per cent
using this technique.
The creation of a neat edge to a wrap around structure can pose
problems. One solution is to use a series of sand or grout bags to
facilitate shaping of the edge. The bags are built up on the inside
of the facing material and act as permanent formwork against which
the fill can be placed and compacted. The necessity for temporary
shuttering is avoided and there are usually no problems of fill
retention. The content of the bags can be constituted so as to
provided a growing medium for grass or quick rooting plants, such as
ivy.
The system has been used in the United Kingdom, and similar
structures in Japan are described by Fukuda, (1986).
2.2.7 Contained Facings. Contained facings are similar in concept
to the sandbag detail, except the facing and support element is
formed using the inherent reinforcing material. Gabions, cubic
containers and pillows fall into this this category. In each case
the facing acts compositely with the reinforcing elements and is
588

Figure 17

Figure 18

geogrid
~.

turf lining

Figure 19
 589

"., Derlor,,~d dnin.pe

pipe to b~ ~"po$~d

.- 10Ga p,l., .t~~1 ... ire

anchor hooks placed.
3'0· Int~,"als ellery other
I.yer.
Aln,U',ATII!fG lAT(IItS Of "('".FQAe";
'.'lIl11e AND CtltuSHEO STON(

t-----It--...~,

CONCU'7E
(F." .... ,

Figure 20

Figure 21
590

usually joined to them.

Reinforced or "tailed" gabion structures developed by Templeman and
Jones (1979) are constructed using gabions connected to horizontal
layers of geogrid reinforcement, Fig. 21. In this system the gabion
facing is erected conventionally and the "tails" are laid in position
and tied to the base of each row. Structures up to 8m in height have
been built in England.
Recent work by LCPC in France has produced a system of construction
using geotextile gabions or "cubic containers/pillows". This is
similar to the "tailed" gabion concept except that the reinforcing
material is a geotextile. This technique was first used to construct
a s . Om high steep stepped faced structure at Trouville Sur Mer,
France, Perrier et aI, (1986).

2.3. CHOICE OF SYSTEM

A number of interacting factors can be shown to influence "the choice
of one system over another and to determine the form of the
construction itself. It is convenient to group the various
influences under three headings:
a) Technical Requirements
b) Market Forces
c) Working Practices
2.3.1 Technical Requirements. The technical elements which
influence the selection, design construction and form of a
reinforced soil structure are shown in Table 1.
It is difficult to determine the degree of importance of separate
elements on the construction system adopted. Tal:>le 2 provides an
indication of potential weightings which might occur, although
individual conditions will produce alternative conditions
(i) Function.
The ultimate use of the proposed structure probably has the maj or
influence on the choice of structure. As an example some materials
can be used to provide the facing as well as the reinforcement
although their appearance is often not acceptable other than for
temporary, industrial or military structures. Similarly, fill
properties and the durability of reinforcing materials may not be
important with temporary contruction but are critical with permanent
structures.
(ii) Subsoil Conditions.
Subsoil conditions are important in all reinforced soil constructions
and often the primary technical reason for the choice of this type of
construction. All of the three basic construction systems can be
used on weak subsoils, although some facings and constructional
details are better than others at accommodating significant
differential settlements. In particular, the wrap-around and gabion
style facings, used with the concertina method of construction, are
able to accept major distortions, Godfrey, (1984). Subsoil
conditions also influence whole body rotations and distortion, a
factor discussed later.
(iii) Rate of Construction.
Construction of reinforced ioil stuctures is normally rapid.
Construction rates of 40-200 m per day per man may be expected and
usually the speed of construction is determined by the rate of
 591

TABLE 1

Reinforcement Soil Construction

Composition Particle Size Construction System

Durability Grading Compaction
Form Index Properties
Surface Properties Mineral content Facing
Dimensions Durability
Strength Availabili ty
Stiffness

-- - - ---

Reinforcement Soil State Structures

Distribution

Location Density Geometry

Spacing Confinement End Use
orientation State of Stress Aesthetics
Degree of Saturation
Drainage
592

TABLE 2
APPLICATIONS

Construct- Temp- Short Perm- Indust- Mili- Archi-

ion orary Life anent rial tary
techtural
Factor
Drain2.ge *** ic** *x* *** *** ***
Distor- * * ** * * **
tion (***) (***)
Subsoil ** ** ** ** *** **
Rate of *** ** ** ** *** *
Construct-
ion
Fill * ** *** * * **
Properties
Reinforce- * ** *** ** ** **
ment
properties
Facing * * *** * ** ***
Aesthetics * * ** * * ***
Durability * * *** ** * ***

* Secondary Importance
** Important
*** Very Important
 593

placing and compaction of the fill. On a recent UK contract the rate

of production of the industrial waste used as fill became of the
limiting criteria on the construction, Jones et aI, (1987). In some
cases the economic rate for the production of facing units may
determine the construction rate, particularly if an original or
unique facing is required. It is essential to determine this aspect
of any project early in the design cycle. Construction is normally
unaffected by weather except in extreme situations.
(iv) Durability and Degredation.
In all situations care should be taken to minimise physical damage to
facing elements and reinforcement, and wherever possible corrosion
must be avoided.
Normally vehicles are restricted so as not to run over exposed
reinforcement and cause damage. Polymeric reinforcement is not
subj ect to corrosion and is usually very durable in conventional
fills, however, it is susceptible to ultra voilet light and must be
stored out of direct sunlight. In some Codes of Practice the
properties of polymeric reinforcing materials are factored downwards
to accommodate possible damage caused by fill/reinforcement
interaction. The use and selection of fill which is benign to
polymeric reinforcement in this context can be advantageous. An
example is the use of pulverised fuel ash as a structural fill with
"Tensar" geogrids, where a material properties reduction factor (r'm)
= <0.9 is taken to be appropriate. At present it is not considered
acceptable to use pulverised fuel ash with metallic reinforcement
because of the risk of severe corrosion.
(v) Aesthetics.
The appearance of most structures is closely associated with use.
With permanent structures appearance is usually of considerable
importance. In the case of structures in urban environments the need
for a high standard of finished appearance can influence the
construction system chosen. In the case of the hybrid system of
construction, masonry faced structures have been found to be the most
economic. Wraparound structures on the faces of which vegetaion is
encouraged can be very attractive, similarly hand placed stone in
gabions has aesthetic appeal.
(vi) DistoJ;tion
Some reinforced soil structures may be prone to distortion,
particulary during construction. Many of the construction details
adopted in practice are chosen to minimize distortion and its
effects.
a) Concertina Construction.
Structures built using the concertina method from geotextiles and
geogrids are particularly prone to distortion of the face. The
degree of distortion cannot be accurately predicted. An accepted
method of overcoming the problem is to cover the face of the
structure either with soil or with some form of facing. An
alternative is to provide a rolling block or sand bag facing against
which the compaction plant can act, to minimise distortions and make
them as uniform as possible, Fig.22.
b) Telescopic Construction.
An estimate of the internal movements and distortion of the facing
can be made from observations of prototype and trial structures.
Typical vertical movements within the soil mass which are
594

74;',
,
I

~'
j
!€;==~~= ___________";I f~bric

reinforcement
IT",.m Rf 12)

~~~==~~------------------~,
I
I

random
masonry
facing

concrete 150Ld.
ellSs 30/20 porous drain Figure 22

'" 10 B

panel
closure
(rnn)
""

J 0
"" '00 300

elevation tine (days)

Vertical IOOvement of panels
,00

tine
"0 \
Figure 23
(days)'OO

'"
O~o~~-7,o~--,~o~--~~--~~--~

panel tilt (rnn)

Variations in panel tilt. (Alter Findlay, 1977)

(0) (bl

Pivot point or racing panels Figure 24

 595

transmitted to the facing are made up of three components:

1. horizontal movements at the connections
2. tilt of the facing units
3. time dependent strains in polymeric reinforcements.
Providing good practice is followed and the reinforcement is slightly
pre-tensioned, movement of connections during construction is not
normally significant and is likely to be 2-Smm. All facing panels in
this form of construction tilt, the pivot point depending upon the
geometry of the facing, Fig. 24. The movement of individual panels
is normally that of an outward rotation (ie. the top of the panels
moving away from the fill). In some special cases the face movement
is determined by whole body rotation which can be a movement of the
structure into the fill. A prediction of the whole body movement
should be part of the design.
c) Sliding method of construction.
Non-Structural Facing
When a non-structural facing is used, distortion of the facing is
likely to occur, the degree being dependent upon compaction.
Distortion can be controlled by:
1. Using light plant in the 2m zone adjacent to the facing.
2. Using bold architectural features to mask the distortion, such
as in Fig. S(a). These can disguise forward rotations and major
bulges by creating a face without a natural sight line.
Structural Facing
When a full height structural facing is used and the construction
method is as in Fig.7, the horizontal movement of the facing will be
limited to the movement capacity provided by the reinforcement/facing
connections and the stiffness of the reinforcing material. Movements
of 2-Smm for walls S-Sm high have been observed with reinforcements
with a high modulus. With more extensible reinforcements, movements
of S-20mm are typical. The facing will normally rotate outwards
about its toe and the movement is usually sufficient to reduce the
lateral soil pressures behind the facings to the active state. Under
some conditions associated with very weak subsoils whole body
rotation of the structure may occur. This may result in a backward
lean of the facing. The propensity of the structure to rotate and
the degree · of rotation is a function of subsoil condition and the
geometry of the structure, it has been suggested that narrow, double
sided (back to back) structures are not prone to whole body
rotations, Jones and Edwards (19S0).
(vii) Logistics.
The speed of construction must be catered for if the full potential
of the use of reinforced soil structures is to be realized. Normally
this will cause little or no problems with the reinforcing materials,
but the production and delivery rate of the facing units may cause
problems, particularly if multiple use of formworks is required for
economy. The delevery of fill has also been shown to be a deciding
factor associated with construction rate.
The cho~ce of structural form and construction technique may depend
ultimately upon the ease and economy of moving constructional
materials. As an example, the light weight of the geotextiles and
related materials makes them suitable for air freight, whereas
concrete and steel elements can have significant transport problems.
596

(viii) Construction Sequence.

Reinforced soil structures encourage the use of non-conventional
construction sequences and technical innovations and may permit a
reduction in the number of construction steps as illustrated in Fig
25.
Alternatively it is possible to change round a construction
sequence as in Fig 26, where the backfill of a structure is placed
before the structure itself. This is achieved by forming the
backfill into a temporary reinforced soil structure and using the
face of this structure as the rear formwork for the permanent works.
This technique has been used sucessfully in bridgeworks construction
in the United Kingdom.
2.3.2 Market Forces. Market forces frequently determine the form
of engineering structures. These forces are not necessarily
synonymous with base or prime cost, but rather with what the market
will bear. At a time when there is general over capaci'ty in the
geotextile industry, cost may fall at times, to levels which are not
necessarily economic.
The development of proprietary materials and systems can be a major
influence on the choice of structure. Development costs are always
large, but through use in a large number of applications the costs
may appear modest, whereas structures specifically tailored with
special one off details and facings can prove very costly. In
addition, the latter may not be well designed. As a result there is
an obvious incentive to duplicate methods and techniques and
rationalise details.
Monopolies in the form of patents have also played a significant role
not only in championing reinforced soil, but also in determining the
form used. Patented systems, usually represent innovation, however,
they can themselves be a hindrance to further innovation in that,
superior elements or systems may be positively discriminated against.
The development of market choice and the acceptance of alternatives
by Commissioning Agencies is the means to eradicate this negative
element.
2.3.3 Working Practices. The civil engineering profession is
inherently very conservative, and innovations are scrutinised with
ri,gour. In addition the life expectancy of most civil engineering
works makes for a cautious approach. As a consequence the
introduction of new techniques and materials can be difficult,
particularly if the use of the new system requires an element of
learning from the user. The development of polymeric products for
use in civil engineering has corne about through market forces,
operating within both the textile and civil engineering industries.
The acceptance of these materials has been due to skilled marketing
within the civil engineering profession rather than as a demand from
the profession.
National Agencies are frequently the key or catalyst to innovation.
In the United Kingdom, developments within the highway field have
occurred largley as a result of the encouragement of the Department
of Transport and the Transport and Road Research Laboratory; without
Technical Memorandum BE3/78 relating to reinforced soil structures,
this form of construction would still be experimental.
 597

st.p3 ,"d 1 /~ ..
~-----

stl!p4

convention.1 s"uctur~ rl'inforCl!d ~jl struc1ure

soil structures

Figure 25 Construction sequence for conventional structures

and reinforced soil structures

Figure 26
598

The greatest encouragement to the use of reinforced soil is the

development of recognised National Design Standards. In the
conservative civil engineering profession Codes of Practice have
become an essential element to many designers. In the absence of a
Code of Practice, acceptance criteria such as the British Board of
Agrement Certificate became essential to progress.
A second influence on the choice of reinforced soil systems follows
from the structure of the professions in different countries. These
differences are important and have been indirectly responsible for
legal difficulties in the reinforced soil field.
The Designer or Consultant is often reluctant to consider
alterations or alternatives to his original design often looking
upon these as a form of criticism. Changing a "conventional" design
to a reinforced soil structure is seen by some designers as very
radical. Even suggesting a change from one form of reinforced soil
to another or the use of a different polymeric produc't can be
resisted. This resistance may be legitimate as the suggestion of an
alternative can often only be made during the tender or bid period.
As acceptance of the alternative places responsibility on the
Designer, resistance to change can be understandable, with the
argument being that all the technical consequences are not known. In
Scotland a method of overcoming the problem of alternative designs
has had some success, Varley (1984).
In the United States, the deliberate action of the Federal Highway
Authority and some State Highway Authorities to encourage competition
has had the desired effect, not only of reducing costs but also of
making possible a wide range of reinforced soil alternatives. The
different structure of the civil engineering industry in Canada and
the United States to that in Europe encourages competition from
specialists each offering their own proprietory system. To ensure
the best choice is made, many Highways Agencies in North America
prefer to provide complete plans for all recognised alternatives.
This permits the Agency to maintain control of the engineering and at
the same time it provides the best suited structures for each
particular site. The extra engineering effort in this approach is
often mitigated by the willingness of many suppliers to provide
complete plans to the Agency, Leary and Klinedinst, (1984).
'Designers themselves can compete successfully in developing
effective and economical reinforced soil structures. In the United
States of America, Georgia State Highway Authority has developed its
own system (GASE, Georgia Stabilised Embankment) based partly upon
the work pioneered by the California Highway Authority, Forsyth
(1978). In the United Kingdom, the West Riding/West Yorkshire
Metropolitan County Council developed sufficient expertise and
experience to tailor reinforced soil structures to specific
situations, Jones (1985).
The lesson from California, Georgia and West Yorkshire is that
expertise develops with application. Reinforced soil is a relatively
recent reintroduction of an ancient technique and experience in the
field is on a steep growth curve. Accordingly the choice of
reinforced soil system or style will be influenced by the level of
expertise available and the past working practices adopted by the
designer.
 599

3. ECONOMICS
The primary advantage gained from the use of reinforced soil
structures may be the improved idealisation which the concept
permits; thus, structural forms which would have been impossible to
contemplate become feasible and economic. It is generally accepted
that the use of reinforced soil walls or bridge abutments will
produce significant savings; costs of (30-50 per cent) relative to
conventional construction.
The savings produced through the use of reinforced soil are
influenced by a range of factors, not least of which is the level of
competition available. Without competition the cost of reinforced
soil is heavily influenced by the cost of the conventional structure
alternative, and savings can be reduced in line with what the market
will bear. Without competition from alternative soil reinforcement
systems costs may be 70 to 80 per cent of conventional structures.
Before embarking upon a reinforced soil structure design an
estimate of the possible cost is required, the most reliable of which
are based upon previous work. The base cost of any reinforced soil
structure is very difficult to determine, in some cases it may only
be determined by the use of a form of direct labour contract.
However, the major factors which influence costs for any practice can
be identified, Table 3

3.1 TOTAL COST

The total cost of a reinforced soil structure is made up of the
following cost elements:
Soil Fill, (Cs)
Reinforcement and connections (CR)
Facing Elements (if required) (CF)
Labour for transport and construction (Ct)
Transport of materials (CT)
Construction (including all ancillary items such as
drainage, copings and facings), (Cc)
Material testing (CMT)
Profit, (P)
Thus total cost,
' TC = (CS + CR + CF + CL + CT + CC + CMT + P) ----- 1
For contractual purposes equation (1) may be reduced to the first
three elements, but with the labour, transportation and the ancillary
costs included:
TC = (C'S + C'R + C'F) --------------------------- 2
Where C's represents the cost of the soil fill, including transport,
placing compaction and material testing
C'R represents the cost of the reinforcement, including transport and
fixing.
C' F represents the cost of the facing, including transport and
erection. Profit is included.
The elements included in equation 1 are inter-related and the
minimum t0tal cost of a structure may be produced by a combination of
the most compatible elements in any particular situation. For
example, if the particular materials testing systems are
unavailable, then the use of reinforcing elements exempt from these
testing requirements may provide the economic solution even if
600

Table 3

Technical Use, design life, quality of detailing,

Factors quality and suitability of fill, testing
requirements, subsoil conditions, parapet
details, analytical models.

Market Competition, size of structure, use of proven

Forces components, familarity of contractor with
reinforced soil techniques logistics.

Working Conforming to Codes of Practice, ,

Practices available specifications, Experience
of designers.
 601

these reinforcing elements are highly priced. Similarly a

combination of construction elements permitting the use of an
indigenous fill or waste fill such as colliery shale or pulverised
fuel ash may be economical. In other cases technical factors may
predominate; in a recent case the connection detail proved to be the
deciding factor between the use of polymeric grid or strip
reinforcement, Jones et al (1987).

3.2. DISTRIBUTION OF COSTS. Using the elemental breakdown given in

equation 2 it is possible to illustrate that the distribution of the
cost elements vary not only with the relative costs of the
constituent materials but also with the dimensions of the structure.
Assuming that the relative cost of the three elements of equation 2
are in the ratios of:
Soil fill, per unit volume (m3) 1.0
Reinforcing elements per unit area of face (m2) 1.5
Facing elements per unit area of facing 10.0
and where the width of the structure is B, the height is Hand (B=H) ,
then the distribution of costs with respect to the height of a
vertically faced retaining wall are shown in Fig.27. It can be seen
that the relative costs of the three basic elements for a 10 m
structure are approximately:
Soil fill 30 percent
Reinforcing Elements 40 percent
Facing elements 30 percent
If the relative costs of the three basic elements of the same
structure are changed to:
Soil fill, per unit volume (m 3 ) 0.5
Reinforcing elements per unit area of face (m 2 ) 2.0
Facing elements per unit of facing 10.0
then, although the total cost of the structure remains the same, the
distribution of the costs is very different, Fig.27.
These data also illustrate the influence that scale or size of
construction may have on costs and show that at lower heights the
influence of the cost of the facing on the overall costs becomes
dominant. With small structures, the material requirements for the
facing may.be of the same order as the material from a conventional
structure, a point reflected in Fig. 28, which also shows the
imprecise nature of cost comparisons. At low heights, particular
attention may be required to reduce the costs of the facing in order
to retain the economic benefit of a reinforced soil structure; one
method known to be successful is to use masonry or brick facing
normally associated with small scale construction or building
techniques. Experience has shown that this form of· structure is
compatible with the use of polymeric reinforcements, Jones (1982).
A second influence on overall cost, associated with the scale of
the project, is the contractural arrangements under which the
structure is built. For some individual structures and structures
under 3 m in height the labour requirements are low and the use of
specialist sub-contractors may prove uneconomic, in which case
economic construction of the soil structure may be attained only by
the main contractor, through local contractors, or by a direct labour
organisation.
602

,. reinforcerrent

. ..
50
50

percentage .. percentage .00

of of
soil fill total
total
cost cost

,.
,. facing -elenents
,.
.+-----r----+----~,----,,-----
o 10'~ 20

height of structure (m)

.
,
.+---~----~----~---­
• os
height of structure (m)

Figure 27

100 , - - - - - - - - - - - ,
econany 7.
100
80

percentage cost 75
of reinforced soil 60
relative to
reinforced concrete
40
cantilever walls

25
20

o+-----~----~----~--~
10 15 20 o 10 15
height of structure (m) height of structure (m)

Figure 28
 603

3.3. CONSTRUCTION. In conventional structures the placing and

compaction of fill may not be associated with the construction. If
reinforced soil structures are treated similarly, the fill element
costs may be removed and the distribution of construction costs
change dramatically. The costs established in one study are as shown
in Fig.29 The volume of structural fill required for use with a
conventional structure may exceed that used in a reinforced soil
structure, Fig. 30.

3.4. COST DIFFERENTIALS. The cost of fill materials is dependent

upon local availability and haulage rates. Similarly, the cost of
facing materials is a function of locality and custom. Reinforcement
materials have different properties and costs vary. Thus, even
though the theoretical cost for different reinforcements may indicate
financial preference for one material, market conditions may give a
different trend. Overriding all considerations is the requirement to
obtain the minimum cost, equation 1.

3.5. ECOLOGY AUDIT. An alternative method of assessing the benefits

of earth reinforced soil systems, is to use an ecology audit. An
advantage of this approach is that it is immune from the commercial
distortions which are associated normally with new constructional
systems, and therefore it may produce a more realistic assessment of
the true costs.
The increase in energy costs has led to an interest in energy
calculations including the energy content of building materials.
However, energy is only one of the ecological parameters needed to
determine the complete effects, (short-term, long-term and
associated) of engineering works. Of growing importance and interest
are the problems created by scarcity of raw materials, the
environmental problems created by pollution, both of the atmosphere
as well as the land from mining activities, the increase in manpower
costs and transportation costs and the cost of maintenance. The
choice of structural form used for any scheme influences all of these
parameters. Determination of the complete costs of a structure to
society may be attempted by studying the ecological parameters
represented - in the whole cycle necessary for its production,
including:
* mining
* raw materials
* process industry
* basic materials
* product/construction industry
* users maintenance
* waste recycling.
In practical terms the ecological parameters associated with a
reinforced soil structures are:
* energy content of the materials forming the structure
* qvantity of process water required to manufacture the
materials.
* despoiling of land necessary to produce the materials
* pollution caused during manufacture and construction.
604

%
Materials Facing 21
Faeing moulds 4
Vertical reinforcements 4
Horizontal rcinforcemcnu 32
Drainage 2
Others 4

67
labour Site clearance I
Retaining wall II
Drainage 5 7

17
PlancJ.and Site clearance I
operatives Retaining wall 10
Drainage 5

16

100

Figure 29 Breakdown of construction costs

H 3m
~I ~
--------~
2 /
""J 1 --:2J 1
,,-/ 2

~3m

volume < H'lm length volume • H2 + 3Hlm length

\'olume of earth fill required in reinforced earth and "inforccd concrete retaining ".lIs

Figure 30 Volume of fill required in reinforced soil and reinforced

concrete retaining walls
 605

* labour costs for material manufacture, transport,

construction and maintenance.
* demolition requirements.
The ecological parameters associated with the construction of
reinforced soil structures formed using reinforced concrete, steel
reinforcement and cohesionless soils are illustrated in Table 4.
Figure 31 shows ecological parameter values for a 64 m prototype
reinforced soil structure compared with an equivalent reinforced
concrete cantilever retaining wall. The reinforced soil structure is
significantly more efficient in ecological terms. Arguably economic
factors have as their ultimate base the ecological parameters;
accordingly Fig.31 is a potent argument that reinforced soil
structures are efficient and economic. In addition, polymeric
reinforcement appears to be competitive when compared with steel,
Table 4.

4. SPECIFICATIONS
Many civil engineering projects worldwide are undertaken by
contract. The system requires that the client and his appointed
representatives have the means with which to adequately design and
properly specify their needs and that the Contractor is fully aware
of these requirements and their implications at the time of tender.
The Specification describes how aspects of the work shall be
executed, referring when appropriate to approved Codes of Practice
and technical guidelines.
The use of polymers as reinforcing materials for soil structures
raises a number of problems with the development of specifications.
Unlike established materials such as concrete and steel, the
properties and characteristics of polymer materials are less well
understood. In addition, because of their relatively recent
introduction and the range of available polymeric materials, standard
tests to evaluate their properties are only now emerging.
For these reasons, the development of Codes of Practice, design
guidelines and contract documentation, for use with polymeric
materials in reinforced soil structures, have tended to divide into
two categories - those for methods and those for materials, indeed in
some countri~s only the former is presently underway.

4.1. METHOD RELATED DOCUMENTATION. There are two main sources of

design and construction information for reinforced soil systems,
(a) Technical Memorandum (Bridges) BE 3/78
"Reinforced Earth Retaining Walls and Bridges
Abutments for Embankments".
Department of Transport (UK) 1978
(b) "Les Ouvrages en terre armee - Recommendations
et reg1es l'art". Laboratoire Central des
Ponts et Chausees. (LCPC), France 1979.
These documents were originally formulated for use with metallic
reinforcement and both have been adopted as the basis for standards
of design 'a nd construction by other countries worldwide. The United
Kingdom, Memorandum BE3/78 caters for the use of non-metallic
reinforcement. However, the increasing availability and use of
polymeric materials has left both documents deficient for the
general case. Accordingly, the tendency has been to develop
606

energy content of construction materials

process water used in manufacture of

materials

despoiling of land in production of

materials

S02 -emission

.just - emission

Idbour - manufacture of matt!rials

labour - material transport

labour - construction

o 20 40 60 80 100 120 140 160

reinforced soil structure
x 100
reinforced concrete structure

Figure 31
 607

Energy consumption of construction materials

--
- -
MATERIAL GJ/ton
Gravel 0.104
Sand 0.128
Ordinary Portland cement (OPC) 8.2
Blastfurnace cement (HC) 3.0
Water 0.004
Mixing concrete 0.058
Mild steel reinforcement (bar, grid) 22.8
Prestressing stell (bar, strand) 28.3
Plastic (high density polythene sheet) 84.0
- - .- - - ._-- - - -- - --- ---- ." - - - - --- - "- --_ . . . - - _ .. ------- --

Concrete (J 4 0 kg Hc/m 3 ~ 2.18 GJ/m J

Concrete (360 kg oPC/m ) 3.28 GJ/m 3
Concrete tiles and bricks 3.18 GJ/m 3

CONSUMPTION OF PROCESS WATER IN MANUFACTURE

Concrete 6.3g 1/m 3

Steel (reinforcement, prestressing, 55 m /ton

DESPOILING FROM PRODUCTION OF MATERIALS

Concrete 0.6~ m2 /m 3
Steel 5 m /ton

POLLUTION - S02 EMISSION

Concrete 0.37 kg/m 3

Steel 2.00 kg/ton

DUST EMISSION

Concrete 1.29 k9/~3

Aggregate/fill (sand, gravel) 1.1 kg/m
Steel 2.7 kg/ton

LABOUR - MATERIAL MANUFACTURE AND TRANSPORT

Concrete 1 man-h/m 3 m~nufacture

1.45 man-him transport
Steel 10 man-h/ton manufacture;
0.6 man-h/ton transport
I
608

guidelines derived from manufacturers recommendations and local

experience.
A number of National documents which cover the use of polymeric
reinforcement are known to exist or be in the course of preparation
including:
(i) "Model Specification for Reinforced Fill Structures"
Geotechnical Control Office, Hong Kong.
ii) "Design Guidelines for the use of Extensible Reinforcements
for Mechanically Stabilized Earth Walls in Permanent
Applications".
AASHTO-AGC-ARTBA, Joint Committee Task Force 27. (Draft 7/86).
(iii)"Guide to Ground Improvement"
Construction Industry Research and Information Association
(CIRIA). UK.
(iv)"Model Specification for the use of Geotextile and Related
Product!;".
Institution of Civil Engineers, London.
(v)"Code of Practice for Soil Reinforced Retaining Structures".
Indian Standards Institution Sub-Committee for Geotextiles,
BDC 23.8 (expected July 1987)
(vi)"Strengthened/Reinforced Soils and other Fills"
UK Code of Practice, British Standards Institution
Committee CSB/56.

4.2. MATERIAL RELATED DOCUMENTATION. There are several different

agencies involved in the development of testing methods and many
countries now have national committees dealing with the development
of test methods for geotextiles and related materials e.g. B.S.I. and
A.S.T.N. and these national test method committees are co-ordinated
internationally through I. S. O. . Several other professional bodies
also attempt to co-ordinate their work, either directly or
indirectly, e.g. R.I.L.E.M., P.I.A.R.C. and I.C.O.L.D.
Progress to date has not been as rapid as first anticipated with
very few test methods agreed which provided data that can be used
directly in design.
For this reason no national or international material
specifications to deal with all aspects of geotextiles and related
materials used as reinforcements is available.
The specification being produced in the United Kingdom. by the
Institution of Civil Engineers is probably the most advanced.

5. DISCUSSION
The number of reinforced soil structures using polymeric
reinforcement is growing. The rate of growth is likely to increase
as the technical and economic advantages of the technique become more
widely appreciated.
From current usage it is possible to identify areas where
improvements can be made, and which, if introduced would assist both
designers and contractors, and also lead to further economies.

5.1. CONSTRUCTION METHODS. The existing construction methods can be

improved, and the following developments have been identified as
likely to be cost effective:
 609

a) The development of standard reinforcement layouts

b) The introduction of more standard facings and details
c) The identification of optimum plant to comply with
construction specifications.
In addition, consideration of the following points may lead
to developments;
d) The need for and effectiveness of pretensioning of
polymeric reinforcement.
e) The logical arrangement of joints.

5.2 ECONOMICS
A better understanding of the economics of reinforced soil structures
is required. Advances in this important area would be greatly
assisted by:
a) The pUblication of case studies which detail actual costs,
including background data.
b) Studies of the long term and maintenance costs, including
further consideration of the durability of polymeric materials.
c) Identification of the factors which influence the economics
of reinforced earth soil structures in different countries.

5.3. SPECIFICATIONS. Many potential designs are being hampered by

the lack of adequate specifications, although the widespread
development of Codes of Practice will resolve many questions. Even
so, it is possible to conclude that:
a) Specifications should be broadened
b) There is a need for standardised tendering procedures
c) There is a neEl~ for accepted methods of assessing different
products and for comparing structures developed using different
polymeric reinforcements.

5.4 FUTURE. Likely future developments will corne from two areas.
The first is through _ the introduction of new advanced polymeric
materials, such as the Directionally Structured Fabrics (D.S.F.)
which have recently been announced.
The second area of future development will result from a general
growth of t.he reinforced soil technique. This will produce more
diverse applications. It is already possible to discern the
development of different "styles" in reinforced soil technology based
upon need and frequently influenced by local topology and local
practice. An example of the latter is the development of the
Norwegian Trondblock walling system to incorporate polymeric grid.
610

REFERENCES

1. Berg, R. R., Boneparte R., Anderson R.P, and Chouery V.E, (1968).
"Design Construction and Performance of Two Geogrid Reinforced
Soil Retaining Walls". Proc. 3rd Int. Conf. on Geotextiles,
Vienna.
2. Cantelli R., and Munfakh G., (1986). "Geotextile Walls in
Mountainous Terrain". Proc. 3rd. Int Conf. Geotextiles,
Viennna.
3. Department of Transport (1978). Reinforced Earth Retaining Walls
and Bridges Abutments for Embankments. Tech. Memo. (Bridges)
BE 3/78, London.
4. Douglas,G.E., (1981). "Design and Construction of Fabric-Rein-
forced Retaining Walls by New York State". Trans Research
Record 872.
5. Findlay, ~.W. (1978). "Performance of a Reinforced Earth
Strucutre at Granton". Ground Engineering 2, No.7, 42-44.
6. Forsyth R.A., (1987). "Alternative Earth Reinforcements". ASCE
Spring Convention, Pittsburg.
7. Fukuoka M., (1986). "Fabric Retaining Wall with Multiple
Anchors". Proc. 3rd Int. Conf. Geotextiles, Vienna.
8. Fukuda N, 'Yamanouchi T, and Miura N., (1986). "Comparative
Studies on Design and Construction of Steep Reinforced
Embankment". Proc. 3rd Int. Conf. Geotextiles, Vienna.
9. Godfrey K.A. Jnr. (1984). "Retaining Walls: Competition or
Anarchy" . ASCE Civil Eng. Dec.
10. Jones, C.J.F.P. (1978). "The York Method of Reinforced Earth
Construction". ASCE, Spring Convention, Pittsburg.
11. Jones C.J.F.P. (1985). "Earth Reinforcement and Soil
Structures". Butterworths, England.
12. Jones C.J.F.P. (1987). "Practical Construction techniques for
Retaining Structures using Fabrics and Geogrids". 2nd Int.
Conf. Geotextiles, Las Vegas.
13. Jones C.J.F.P., Murray R.T., Temporal J. and Mair R.J. (1985).
"First Application of Anchored Earth". XI ISSMFE San
Francisco, August.
14. Jones C.J.F.P., Jamison W, and Garner D. (1987). "Design
Construction and Economics of Dewsbury Retaining Walls". In
print.
15. Jones C.J .F.P., and Edwards L.W. (1980). "Reinforced Earth
Structures Situated on Soft Foundations". Geotechnique, June.
16. Kempton G.T, Entwistle, R.W. and Barclay M.J. (1985). "An
anchored fill harbour wall using synthetic fabrics". Proc.
Inst. Civil Eng. Part 1, vol.78 April pp 327-347.
17. Laboratoire Central des Ponts et Chausees (1979). "Les Onulages
en terre armee - Recommendation et regles d'art". Paris.
18. Leary R.M., and Klinedinst G.L., (1984) "Reinforced Wall
Alternatives". FHWA, Washington.
19. Perrier, H., Blivet, J-C, and Khay, M, (1986). "Experimental and
,Actual use of Geotextile Reinforcement of a Slope". 3rd Int.
Conf. Geotextiles, Vienna.
20. Pigg D.Z., and McCafferty W.R., (1985). "The Design and
Construction of Reinforced Soil Retaining Wall at Low
Southwick, Sunderland". Polymer Grid Reinforcement, Thomas
Telford Ltd.
 611

21. Saran, S., Talwar D.V., and Prakash, S. (1979). "Earth pressure
distribution on retaining walls with reinforced earth
backfill". C.R. Col. Int. Renforcement des Sols, Paris.
22. Templeman J., and Jones C.J.F.P. (1979). "Soil Structures using
high tensile plastic grids". U.K. patent No. 7941627.
23. Varley W.R., (1979). "Aspects of Alternative Tendering". BCSA
Nat. Struct. Steel Conf. Part 2, London.
24. Vidal, H (1978). "The development and future of Reinforced
Earth". Keynote address. ASCE, Spring Convention, Pittsburg.
25. Vidal, H. (1978). "La terre armee". Annales de L'Institut
Technique du Batiment et des Travaux Publics, 19, Nos. 223-224,
July-August, France.
26. Yamanouchi, T (1986). Private Communication.
 REVIEW OF SESSION

CONSTRUCTION METHODS AND ECONOMICS

Jones opened this session with a general appreciation of the many factors
affecting the economics of geosynthetic construction. Apart from
regional market force aspects, which in many cases control cost and
methodology, he pointed out that working practices may be the dominant
cost factor. The actual cost of the reinforcement is in most cases a
small proportion of the total cost of the structure and as such may have
little influence on overall economics. He then discussed the concept of
using an "Ecology Audit" to establish comparable base costs for different
forms of reinforced soil. An ecology audit sums the overall cost to
society of a particular form of construction especially with respect to
the energy used throughout the process. It therefore avoids to a great
extent market factors and other artificial variables. Such an audit
indicates that geosynthetic reinforcement is a cost effective means of
construction. Despite this economic advantage designers are still hesi-
tant to use this form of construction due primarily to lack of experi-
ence. Therefore the keys to market growth are Specifications and Codes of
Practice for geosynthetic reinforced soil as such standards relieve the
"tension" from the design situation by providing set terms of reference.
Work is progressing in a number of countries on these standards.

The focus of the session then changed to discuss variations in construc-

tion methods. Jones presented a series of slides showing aspects of
construction for a number of reinforced soil structures. Many interesting
questions, comments and points of amplification were contributed during
this stage of the session. The following sections attempt to summarize
the topics discussed:

The advantages and methodology of providing horizontal and vertical

compliance in the connection between the reinforcement and the facing
element were addressed by Bonaparte, Christopher and Jewell. The
method of connection is vital as it is a construction detail often not
considered in the basic earth pressure design calculations. Problems
may arise if a rigid face and rigid connection exist and vertical
settlement occurs in the soil behind the wall. Reinforcement failure
at the connection and even crushing of the facing have been observed
in such circumstances as the vertical forces caused by the settling
backfill far exceed the lateral earth pressure derived forces for
which the reinforcement had been designed.

The problems of bidding for jobs, where the fill to be used is tightly
specified to allow steel reinforcing to be used, when in fact much
cheaper fills could be safely used with geosynthetics . This problem
ment ioned by Delmas, Bonaparte, McGown and Christopher pointed once
again to the need for Specifications and Codes specifically for
geosynthet ics. Jones pointed out that the maj ori ty of the large

613

P . M. Jarrett and A. McGown (eds.). The Application of Polymeric Reinforcement in Soil Retaining Structures, 613-615.
© 1988 by Kluwer Academic Publishers.
614

number of case studies that he had presented could in fact not have
been constructed using steel reinforcement due to the corrosive nature
of many of the fills.

One aspect of marketing that found general consensus was the

impression that where good technical design support was available from
particular manufacturers it does seem to bear fruit in terms market
share.

The provision of architecturally acceptable facings was identified as

a vital need for geotextile walls by Leflaive.

It was pointed out by Jones that strong, inextensible materials such

as "Paraweb" were available and had been successfully used on struc-
tures up to 28 m high. In addition similar yet stronger materials
could be dev-eloped using Kevlar. One slide of the ziggurats in
ancient Babylon left Christopher musing about the possibilities of
performing wide width tensile tests on the reed mats from which they
were constructed and Jones hoping that one day he will build something
as high that lasts as long.

At the end of this session, formal presentations were made by four

speakers. Bush discussed a case history in which a void had formed around
the connection between a facing unit and the reinforcement. He used this
case to make the cautionary point that the "in-isolation" strength of the
reinforcement is the weakest state and that care must be taken if one were
to use higher strengths from "in soil" tests to ensure that the construc-
tion procedures do in fact give "in soil" conditions. Leflaive discussed
a case history of a motorway fill that was constructed with a temporary
vertical face up to 8 m high using a woven polyester geotextile and no
formwork. The face itself was stabilized using Texsol. McGown discussed
results obtained in a 1 m high testing facility that allows varying
degrees of horizontal stiffness at the wall facing elements. He concluded
by indicating that greater soil resistance is mobilized as larger horizon-
tal movements occur leaving less force to be resisted by the wall or the
reinforcement. In practical terms the movements necessary for development
of the Active earth pressure state may be encouraged by limiting the
backfill' compaction immediately behind the wall or by slackness in the
facing to reinforcement connection. However he suggested that one may
more effectively guarantee and control such compliance at the face by
providing a thin compressible polymeric drainage layer directly next to
the facing. Finally Milligan discussed the soil movements and strains
that occur behind retaining walls as the walls move or rotate. He indi-
cated that for reasonable practical levels of wall movement that horizon-
tal soil strains of approximately 1% would occur and that such levels of
strain represent the probable working strains for the reinforcing
materials.

In summation this session stressed the urgent need for Standard Codes and
Specifications directed at geosynthetic reinforced soils. A major benefit
from such standards will be the allowable use of cheaper, lower quality
fills with geosynthetic reinforcement than are presently allowed with
metallic reinforcement. At present many authorities in the absence of
 615

such Codes insist on having geosynthetic reinforcement bid directly

against metallic reinforcement based on the use of a common high grade
backfill. In terms of design detail both the stiffness of the facing and
the compliance of the connection between the facing and reinforcement must
be considered during design. The connection can be a point of stress
concentration if the backfill settles. The stresses developed in a rigid
connection may easily exceed those estimated from lateral earth pressure
considerations.
Research Needs
RESEARCH NEEDS, REVIEW OF SESSION

DR . K. ANDRAWES, University of Strathclyde, Scotland

DR. P.M. JARRETT, Royal Military College of Canada

1. INTRODUCTION

This chapter is intended to identify the areas which require further work
and research in order to improve the understanding of geosynthetic -r ein-
forced soil behaviour. From such enhanced comprehension should come the
establishment of safe yet more economic design methods and further
advances of the construction technology for such structures. The material
presented is based primarily on the extensive discussion which took place
in the final session of the workshop and which involved all those partici-
pating. In addition various speakers and authors during the workshop
identified topics requiring further study. These too have been included.
The statement therefore represents to some extent the product of the
research workshop as it records the questions still unanswered and felt by
the participants to be of importance.

To provide a framework for the presentation, the needs will be presented

under the following five main headings even though it is realized that the
areas are inter-related and some repetition may arise:

Testing soils and reinforcements

Testing systems
Design
Construction
Developments of materials and techniques
,
One final aspect of introduction is to point out that there are two
distinct levels of needed research. Initially and of greatest urgency is
the need to provide information that will allow the development of
accepted, simple, day to day design methods that will in appropriate
circumstances encourage the use of geosynthet ic reinforcement in soils.
Secondly there is the need to provide a fully developed rational
explanation of the behaviour of geosynthetic reinforced soils. In
suggesting the research needed for this second goal it is most important
that it be realized that the initial goal can be reached prior to
completion and even without many aspects of that work.

2. TESTING SOILS AND REINFORCEMENTS

2.1 Testing Soils

(1) There is an urgent need to def ine the environmental working condi-
tions for geosynthetic materials in the field. The following soil
descriptions are needed for a complete range of soils and conditions
likely to arise in backfills:

619

P. M. Jarrett and A . McGown (eds.), The Application of Polymeric Reinforcement in Soil Retaining Structures, 619-625.
© 1988 by Kluwer Arlldemic Publishers.
620

(i) Chemical environment

( ii) Biological environment
( iii) Temperature regime.

The first two conditions are needed in order to define the parameters
for which the durability of the polymeric materials should be assessed.
The third aspect will set representative temperature ranges to be
considered in design and polymer testing. Of special interest is the
possibility of. high temperatures directly behind facing elements.

(2) Stress-Strain relationships: For most simple design, the direct

shear box may be used to define the value of • and the interface friction
between the geosynthetic and the soil. However for more accurate
prediction of reinforced soil behaviour, knowledge of the manner in which
strains develop, espec ially horizontally, is important. Unfortunately
conventional apparatus, such as the shear box and triaxial test do not
give acceptable relationships involving strains in the soil, especially
post peak. Hence it may be necessary to use the simple shear or other
apparatus to gain better information on strain behaviour. This research
need for a better comprehension of strain development in soils has long
been a fundamental and rather intractable problem for soils analysts.

2.2 Testing Polymeric Materials

(1) The most urgent need is the development of comprehensive durability

testing methods. These should assess the potential change in properties,
especially strength properties, caused by the following effects:

(i) Chemical
(ii) Biological
( iii) Ultra violet
(iv) Ageing under stress
(v) Construction damage
(vi) Temperature and humidity

(2) Stress-Strain relationships: The simplest method for determining

these re,.lationships is to conduct "in- isolation" tests under constant
temperatures that are relevant to field conditions. More data regarding
creep, fatigue and dynamic behaviour are required. Ideally however these
properties should be measured using "in-soil" tests. This represents a
far greater undertaking and so for simple work it would be useful if a
relationship between in-isolation and in-soil behaviour could be
established for the major groupings of geosynthetic reinforcements.

(3) Communications with Polymer and Textile specialists: The range and
complexity of polymeric materials available as reinforcement is continu-
ally increasing. Civil engineers researching the use of these materials
must learn more about their most basic properties and methods of manufac-
ture. There is therefore a serious need to develop communications and
cooperative research with polymer and textile specialists especially with
regard to the durability problem and the development and design of new
products. Please note that this is not the usual pious platitude but a
very real and serious need! Civil engineers must have dialogue with
specialists who can assist them in dealing with such questions as the
effect of chemicals on the behaviour of different polymers, what environ-
mental conditions should be avoided, is biological attack most likely
 621

under anaerobic or aerobic conditions, etc. In fact the dialogue is vital

in terms of even defining what questions should be asked concerning the
long term behaviour of polymers.

2.3 Testing Composite Materials

(1) A better understanding of the composite behaviour and the effects of

interaction between the constituent parts of mUlti-layer soi1-geosynthetic
systems is required.

(2) Very little information exists concerning the reinforcement of fine-

grained soils. Data is required concerning frictional development between
the geosynthetic reinforcement and such soils. Frictional development is
dependent on the pore water pressure regime in the compacted backfill.
When thick geotexti1es that allow inp1ane drainage are used i~ clays , then
rapid consolidation may occur near the geotexti1e leading to an improved
strength and frictional resistance. This interactive aspect requires
investigation.

(3) Pullout testing can be a useful method of estimating soil-geosyn-

thetic interaction especially with regard to anchorage. Standardization
of test size and methodology is needed however to enable meaningful
exchange and comparison of information.

(4) Testing composite materials can mean testing samples of soils con-
taining polymers with each sample considered as a unit. The results
obtained are not divided between the constituent materials but refer to
the unit as a whole. Such an approach was used in the early years of the
development of soil reinforcement, and although many researchers have
abandoned this technique in favour of testing the constituent materials
separately, the composite testing approach may still offer advantages in
certain situations. Recently there has been a revival of this type of
testing due to the introduction of new methods of soil reinforcement using
randomly oriented polymers as in Texsol and the mesh element technique.
It seems that a large scale triaxial apparatus is most suitable for the
determination of the properties of the composites. However, other types
of testing &hou1d be explored and the parameters relevant to design should
be identified.

3. TESTING SYSTEMS

Systems is meant to infer complete structures in which the reinforcements

are embedded in the soil and the behaviour of the structure and its compo-
nents are investigated under both self weight and external loading. Such
tests play a crucial role in enhancing the understanding of the interac-
tion between the soil and the reinforcements, examining the validity of
design assumptions and exploring the effects of various parameters on the
overall behaviour. Systems may be scale models or full sized structures.

(1) There is a need for all scales of systems to be tested preferably

with a progressive development of the scale up to full size. The problems
of scale and time effects should not be overlooked.

(2) Centrifuge testing is also proving to be a useful means of modelling

and should be further explored.
622

(3) For all scales of systems tested, measurements of stresses and

strains should be made in both the soil and the reinforcement and the
effects on behaviour of boundary conditions and construction techniques
should be assessed. For scale models, the testing of an unreinforced
system may often help in analysing and isolating boundary effects.

(4) Parametric studies in physical or numerical model tests are needed to

explore the effects of the stiffness of the foundation soil, the stiffness
of the facing, the stiffness of the reinforced soil mass and spacing of
the reinforcement relative to soil size.

(5) Earthquake and dynamic loading of re inforced soil structures should

be examined and should include consideration of likely pore water pressure
effects.

(6) Stiffer and stronger geosynthetic reinforcements are being developed

and should allow the construction of higher walls. These should be evalu-
ated.

(7) Full scale tests or case histories are, as usual, of inestimable

value . Very few fully documented case histories have been published and
so there is an urgent need to fully instrument and report the behaviour of
full scale structures. Factors to be studied include:

(i) Stresses and strains in the soil both horizontally and ver-
tically
( ii) Stresses and strains in the reinforcement
( iii) Forces at the reinforcement to facing connection
(iv) Vertical earth pressures to confirm any variation from uni-
form overburden pressure
(v) Movements of the wall facing
(vi) Temperatures throughout the , backfill

(8) Of specific interest from field structures besides their general

performance is information on:

(i) Damage caused to the reinforcement during construction

(ii) Effect of damage to facing elements on the overall stabil-
ity . Such damage may arise from fire, vandalism or traffic
impact.
( iii) Long term creep and durability of the polymeric materials.

(9) It is most desirable to assemble a well documented data bank of case

histories of both successful structures and, if they occur, failures. The
case histories should be prepared to some standardized format. The
information tabulated by Yako and Christopher in this volume is a useful
starting point for the scope of information required.

4. DESIGN

4.1 General Problems

The design of soil reinforcement systems offers a great challenge to geo -

technical engineers due to the numerous inter-related factors involved.
The obvious problems of a general nature that require attention are:
 623

(1) Lack of a suitable definition for failure of the system. For

extensible polymeric reinforcements failure should take into account both
servicability criterion for excessive movement and collapse.

(2) What in fact do the factors of safety of reinforced soil structures

and the partial factors of safety of the constituent materials mean? How
realistic are the Factors of Safety presently being used? Factors of
Safety must be defined starting from the basic premise that F 1
represents a failure condition. To enable this to be achieved much more
needs to be learned about failure in these structures.

(3) Lack of clear design assumptions, concepts, specifications and guide-

lines which are applicable to all situations and take into account all the
parameters involved.

(4) Choice of soil and reinforcement parameters to be used in design.

(5) Distinction between the design of a safe structure and the prediction
of its behaviour in terms of deformations and stresses.

Although real progress in design technique will only be achieved by solv-

ing these problems, it seems advantageous at this stage of development to
consider two levels of design, simplified design and rigorous design.
The former will allow the construction of safe simple structures using an
easily understood design methodology and at the same time the latter will
permit researchers to explore complicated concepts in order to achieve
coherent and valid rigorous techniques.

4.2 Simplified Design

(1) For small uncomplicated structures this should be based on existing

limit equilibrium methods which are, in general, safe and very conserva-
tive. The soil and reinforcement parameters should be obtained from
simple tests and used in an easy to understand design procedure. In most
cases the use of <l>cv' obtained from shear box tests, and the assumption
of active earth pressures seems appropriate for the soil except when
relatively inextensible polymeric materials are employed. The design
should allo~ for the calculation of the global factors of safety and the
approximate determination of the working strains.

(2) Design methods are needed to consider Earthquake and Dynamic

Loadings.

4.3 Rigorous Analysis and Design

This can be performed using various techniques such as a discrete consti-

tuent materials approach, a system simulation approach, an equivalent
composite material approach and finite element models. Preference for
rigorous analysis is for "transparent" mathematical models that logically
relate test results for the basic soil and polymeric materials to the
performance of the structure.

In addition to the general problems mentioned earlier, rigorous analysis

should consider and encompass the following:

(1) Calculations and relevance of earth pressure coefficients.

624

(2) Strain development in the soil including strain anisotropy.

(3) Compatability of strains between the soil and the reinforcement.

(4) Effect of reinforcement extensibility and spacing.

(5) Stiffness of the reinforced soil mass, of the facing elements and of
the foundation soil.

(6) Prediction of wall movements.

(7) Effects of creep and relaxation.

(8) Analysis of complex structures including three dimensional problems.

(9) Simulation of compaction induced stresses.

(10) Simulation and allowance for construction detail and procedures,

ensuring that the assumed design boundary conditions are reasonable.

5. CONSTRUCTION

A wide variety of construction techniques are presently being employed,

and it is felt that these would benefit from a greater measure of stan-
dardization and general specification. Work should be directed towards:

(1) Preparing a comprehensive summary of details and variates of differ-

ent existing construction methods.

(2) Identification of desirable and undesirable practice and establishing

the limits of applicability of polymeric products.

(3) Establishing a number of standard forms of construction . These would

include specifications for standard reinforcement layouts, standard
facings and connections and model construction practice and equipment.

(4) Examination of the effects of construction methods and construction

details en boundary conditions and the assumptions used in the design
process.

Standards, Codes of Practice and Specifications created directly for

polymerically reinforced soil structures are vital for the expeditious
development of this promising form of construction.

6. DEVELOPMENT OF MATERIALS AND TECHNIQUES

The success of polymer reinforced soil systems may be attributed in part

to the imagination and development work carried out by the workers in this
field . Although complete understanding of the system behaviour has not
yet been achieved, development work must continue in order to ensure safe
and economical design. Such work should include:

(1) Development of new composite materials, such as Texsol and mesh

element reinforced soil.
 625

(2) Development of methodology and techniques necessary for the use of

low cost, fine grained back-fills.

(3) Development of aesthetic facing units and methods of attaching them

to the reinforcements.

(4) Development of construction techniques to both minimize and allow for

post construction deformations.

(5) Interaction of civil engineers with polymer experts and chemical

engineers for the understanding and development of new materials
especially stiffer and stronger products.

(6) A forum to act as a focus and coordinating body for international

research and development efforts. This could be organized through IGS or
ISSMFE. It should encourage further workshops of the nature of this NATO
meeting, cooperative research projects and the establishment of a data
bank of case histories and other information necessary for the progressive
development of polymerically reinforced soil structures.
Participants
628

PARTICIPANTS

NATO-ARW, "Application of Polymeric Reinforcement

in Soil Retaining Structures"

June 1987, Royal Military College of Canada, Kingston

 629

LIST OF PARTICIPANTS

(1) H. Miki (19) G. Werner

(2) J. Smith-Meyer (20) J. Perfetti

(3) P. Delage (21) E. Leflaive

(4) M. deGroot (22) R. Bathurst

(5) J. Scott (23) C. Jones

(6) R. Berg (24) F. Schlosser

(7) K. Andrawes (25) Ph. Delmas

(8) B. Christopher (26) M. Sot ton

(9) G. Richardson (27) S. Christoulas

(10) R. Bonaparte (28) R. Koerner

(11 ) G. Raymond (29) R. Gopal

(12) J. Lafleur (30) K. Rowe

(13 ) S. Hermann (31) R. Murray

(14) R. Floss (32) A. Rollin

(15) P. Jarrett (33) A. McGown

(16 ) J. Studer (34) P. Rimoldi

(17) D. Bush (35) J-P. Gourc

(18) R. Jewell (36 ) D. Cazzuffi

Missing from photograph:

G.W.E. Milligan J.M. Rigo

630

ADDRESS LIST OF PARTICIPANTS, ARW 588/86

Dr. K. Andrawes Dept. of Civil Engineering, Strath-

(Kamal) clyde University, 107 Rottenrow,
Glasgow, G4 ONG Scotland

Dr. R.J. Bathurst Dept. of Civil Engineering, Royal

(Richard) Military College of Canada, Kingston,
Ontario, Canada, K7K 5LO

Mr. R.R. Berg The Tensar Corporation, 1210 Citizens

(Ryan) Parkway, Morrow, GA, 30260, U.S.A.

Dr. R. Bonaparte GeoServices Inc., Suite 204, 1200 S.

(Rudy) Federal Highway, Boynton Beach, FL,
33435, U.S.A.

Dr. D.I. Bush Netlon Limited, Kelly Street, Mill

(David) Hill, Blackburn, Lancashire BB2 4PJ
England

Mr. D.A. Cazzuffi ENEL - Centro Ricerca Idraulica, e

(Daniele) Strutturale, Via Ornata 90/14,
I-20162 Milano, It~ly

Mr. B. Christopher STS Consultants, 111 Pfingsten Road,

(Barry) Northbrook, IL, 60062, U.S.A.

Dr. S. Christoulas Director Soils Division, Research

(Stavros) Center of Public Works, Secretariat
of Public Works, Ymittou 81, 155 62
Holargos, Greece

Ir. M.T. de Groot, M.Sc. Delft Geotechnics, Stieltjesweg 2,

(Max) P.O. Box 69, 2600 AB Delft, The
Netherlands

Dr. P. Delage Ecole National des Ponts et

(Pierre) Chaussees, CERMES, Central 2 La
Courtine, Boite 105, 93194 Noisy Ie
Grand Cedex, France

Dr. Ph. Delmas Laboratoire Central des Ponts, et

(Philippe) Chaussees, 58 boulevard Lefebvre,
75732 Paris Cedex 15, France
 631

Professor Dr.-Ing. R. Floss Ordinarius fUr Grundbau, Boden-

(Rudolph) mechanik, und Felsmechanik,
Technischen Universitat MUnchen,
Baumbachstrasse 7, 8000 MUnchen 60,
West Germany

Dr. J-P. Gourc IRIGM, Universite de Grenoble, BP 68,

(Jean-Pierre) 38402 St. Martin d'Heres Cedex,
France

Mr. S. Hermann Norwegian Geotechnical Institute,

(Steinar) P.O. Box 40, Taasen, 0801 Oslo 8,
Norway

Dr. P.M. Jarrett Dept. of Civil Engineering, Royal

(Peter) Military College of Canada, Kingston,
Ontario, Canada, K7K 5LO

Dr. R. Jewell Dept. of Engineering Science,

(Richard) University of Oxford, Parks Road,
Oxford OXI 3PJ England

Professor C.J.F.P. Jones Dept. of Geotechnical Engineering,

( Colin) Drummond Building, University of
Newcastle Upon Tyne, Newcastle Upon
Tyne NEI 7RU England

Dr. Robert M. Koerner Dept. of Civil Engineering, Drexel

(Bob) University, Philadelphia, PA, 19104,
U.S.A.

Dr. J. Lafleur Ecole Poly technique , CP6079, Station

(Jean) A, Montreal, Quebec, Canada, H3C 3A7

Dr. E. Leflaive Laboratoire Central des Ponts et

(Etienne) Chaussees, Direction des Programmes
et Applications, Orly Sud No. 155,
94396 Orly Aerogare Cedex, France

Professor A. McGown Dept. of Civil Engineering, Strath-

(Alan) clyde University, 107 Rottenrow,
Glasgow, G4 ONG Scotland

Mr. H. Miki Dept. of Civil Engineering, Drexel

(Hiroshi) University, Philadelphia, PA, 19104,
U.S.A.

Dr. G.W.E. Milligan Dept. of Engineering Science,

(George) University of Oxford, Parks Road,
Oxford, OXI 3PJ England

Dr. Richard T. Murray Transport and Road Research Labora-

(Dick) tory, Old Wokingham Road,Crowthorne,
Berkshire, RGII 6AU England
632

Mr. J. Perfetti Rh8ne Poulenc Fibres, Departement

(Jacques) Nontisse BIDIM, 44 rue Sa'lvador
Allende, B.P. 80, 95871 Bezons Cedex
France

Dr. G.P. Raymond Dept. of Civil Engineering, Queen's

(Gerry) University, Kingston, Ontario, K7L
3N6, Canada

Dr. G. Richardson Soil and , Material Engineers, Inc.,

(Greg) 1903 Harrison Avenue, Box 609, Cary,
North Carolina, 27511, U.S.A.

Dr. J-M. Rigo Civil Engineering Material Depart-

(Jean-Marie) ment, Civil Engineering Institute,
University of Liege, Quai Banning
6-4000 Liege, Belgium

Ing. P. Rimoldi Technical Director, Civil Engineering

(Pietro) Applications, RDB Plastotecnica
S.p.A., 22060 Vigano Brianzi (CO),
Italy

Dr. A.L. Rollin Ecole Poly technique , CP 6079, Station

(Andre) A, Montreal, Quebec, Canada, H3C 3A7

Dr. R.K. Rowe Geotechnical Research Center, The

(Kerry) University of Western Ontario,
London, Ontario, Canada, N6A 5B9

Professor F. Schlosser Terrasol, Bureau d'Ingenieurs-

(Francois) Conseils en Geotechnique, Tour
Horizon, 52, quai de Dion Bouton,
92806 Puteaux Cedex, France

Dr. J.D. Scott Dept. of Civil Engineering, The

(Don) University of Alberta, Edmonton,
Alberta, Canada, T6G 2G7

Mr. J. Smith-Meyer Konglungveien 143, N-1392 Vettre,

(Johan) Norway

Dr. M. Sotton Directeur de l'Institut Textile, de

(Michel) France Section LYON, Avenue Guy de
Collongue, BP 60, 69132 Ecully Cedex,
France

Dr. J. Studer GSS Consulting Engineers Ltd.,

(Jost) Witikonerstrasse 15, CH-8032 Zurich,
Switzerland

Dipl.lng G. Werner Chemie-Linz AG, St. Peterstrasse 25,

(Gerhard) 4021 Linz, Austria
Index
 INDEX

Analysis, 365-407, 409-457, 459- geotextiles, 262, 263, 265,

506, 507-540, 561-567, 569- 287-310, 332-334
570 geogrids, 71-125, 260-261,
circular arc, 429, 463-464, 263-265
488-490 A7 Motorway, France, 303-306
coherent gravity, 107, 423 A15 Motorway, France, 288-291
Coulomb wedge, 427, 429, 468, Cascade dam, 264, 276, 279
564 Gaspe sea wall, 264, 273, 277,
discrete materials approach, 278
344, 543 Glenwood canyon wall, 263, 273
displacements method, 476-481, La Houpette embankment, 291-296
496, 500 Langres structure, 296-300
equivalent composite materials Lithonia wall, 273, 275, 577
approach, 345, 542 Lixing structure, 300-303
finite element, 130-134, 320, Rouen walls, 60, 296, 307, 308
423, 426, 541-553, 557-560 Tuscon walls, 264, 273-274, 419,
ideal length reinforcement, 583
371-374
log-spiral, 429, 465 Class A predictions, 130-157
Rankine, 135, 268, 271-276,
431-433 Classification
system simulation approach, 345 analytical approaches, 344-346
truncated length reinforcement, reinforcements, 30, 32, 410, 411
374-379 systems, 4
two part ~edge, 429-430, 433- testing of polymers, 347-349
437, 443, 466-468, 483-487
Connections, 108, 355, 444, 574-
Analytical methods-comparisons, 595, 613
367-387, 431-433, 459-506
Construction methods, 33-41, 573-
Anchors, 4, 6, 7, 16 598, 608, 609, 624
anchored earth, 6, 7, 263
Backfill, 267, 285-287, 306-308, concertina, 574, 575
619-620, 625 contained facings, 587
clayey, 288, 306, 621, 625 hybrid, 577
gravel, 297, 303 incremental, 75, 581
sand, 3-01 ladder walls, 4-6
silty, 296, 307, 308, 625 propped full height, 75, 581
sliding, 36, 574, 578
Boundary conditions, 402-404 telescope, 574, 576
Tronderblock, 583
Case histories, 239-280, 287-309, Websol, 38
323, 332-334, 622
635
636

wrap around, 262-263, 332-334, gabions, 33, 35

583 galvanized steel, 12, 24, 25,
York, 36 581
geogrid, 35, 303
Construction sequence, 287-306, shotcrete, 28
387-391, 425 structural, 298, 334, 595
wrap around, 262-263, 288-309,
Creep, 51-57, 119-122, 308, 319, 332-334
320, 353, 620
Factor of safety, 268, 319, 444,
Direct shear tests, 19, 22, 112- 483-504, 507-540, 565, 623
115, 181-184, 268, 354, 359-
361, 368, 394, 620 Finite element analysis, 130-134,
320, 423, 426, 541-553,
Drainage, 306-308, 621 557-560

Durability, 19, 32, 33, 57-60, 346, Internal stability, 317-319, 433,
354, 593, 620 508-530
ageing, 308, 309, 620
biological, 620 Joints, 108, 355, 444, 574-595, 613
chemical, 309, 620
construction damage, 173-180, Ladder walls, ~-6
292, 309, 535, 620
ultra-violet, 58, 620 Limit states, 314-316, 326-328,
381, 382, 427
Dynamic behaviour, 620, 622
Paraweb, 38, 581
Earth pressures
Meyerhof distribution, 135, 146, Participants, 628-632
273-275, 423
vertical, 106, 143, 146, 162, Prediction symposium
167, 186-187, 227, 273, 324, class A predictions, 130-157
330, 475 comparison to measurements,
157-167
Ecology audit, 603, 605-607 discussion, 168-170
RMC tests, 71-125
.
Economics, 268-270, 599-605, 609
Properties
Extensibility of reinforcement, 31, geogrids, 118-122, 173-180, 203-
42, 286, 287, 409-453, 507- 205, 401, 447
508 geotextiles, 45-57
polymer, 41-45, 51-52, 620-621
External stability, 315-317, 443, soil, 392-400, 411-415, 619-620
460-461, 530-534 testing, 346-355, 619-621

Facing movements, 88, 89, 96-98, Pullout tests, 16, 18, 268, 355,
141, 156-158, 164, 229, 230, 359-361, 621
290, 294, 320, 330-331, 382-
391, 593-595 Reinforced earth, 3, 7-12, 32, 285-
287, 429, 581
Facings, 267, 418, 625
armateer, 35 Reinforced soil, 3
concrete, 12, 35, 38, 426, 581
contained facings, 587 Reinforcement types, 30-33, 266
fabric, 33, 587 fiberglass, 5, 32, 581
fiberglass, 36 fibers, 19-24, 38-41, 340
 637

galvanized steel, 5, 262 Strains in reinforcement, 91-95,

geogrids, 17, 36, 71, 264-265, 100-104, 160-161, 165-166,
426 227-228, 264, 271-279, 291,
geotextiles, 12, 33, 262, 265, 294, 297-299, 302, 305, 318,
288-310, 332-334 323, 439-443
grids, 14
Paraweb, 38, 581 Stress paths, 393-395, 411-415
stainless steel, 5
strips, 262, 265 Tensar
welded bar mat, 14, 265, 426 properties, 118-122, 173-180

Research needs, 619-625 Tensile testing, 350-355

in isolation, 47
Royal Military College walls, in soil, 45
71-125 strain rate, 49-51, 353
analysis by Bonaparte, 445-453 width of specimen, 47, 48, 352
analysis by Gourc, 496-497, 500
analysis by Jewell, 193-235 Terre Armee, 285-287, 459
geogrid properties, 118-122,
174-178, 203-205, 447 Tervoile system, 24-26
instrumentation, 84-86
sidewall friction, 72, 76, 181- Texsol, 19, 21, 38-41, 340, 345
191, 205, 214-225, 404
soil properties, 76, 77, 112- Triaxial testing, 9, 22, 116, 117,
117, 181-184, 189, 191, 621
195-202, 447
test description, 77-83, 185 Websol, 38
test facility, 71-75
test results, 87-108, 112-125
unreinforced wall, 185-191

Sidewall friction, 72, 76, 135,

181-191, 205, 214-225, 404

Soil behaviour, 195-202, 392-400,

411-415
,
Soil nailing, 26-29

Soil-reinforcement interaction, 380

Seismic design, 502, 620, 622

Specifications, 605, 608, 609,

613-615, 624

Stabil ity, 308

soil behaviour, 195-202, 392-
400, 411-415

Strain compatability, 231-232, 392,

393, 406, 415-425, 522-530,
569, 624

Strain gauges, 84, 86, 271, 289,

292, 297, 301, 303