Engineering Treatment of Soils

Engineering Treatment of Soils
OTHER TITLES FROM E & FN SPON

Cyclic Loading of Soils
Edited by M.P.O’Reilly and S.F.Brown
Drainage Design
Edited by P.Smart and J.G.Herbertson
Durability of Geotextiles
RILEM
Foundations on Rock
D.C.Wyllie
Geology of Construction Materials
J.E.Prentice
Geology for Civil Engineers
A.McLean and C.Gribble
Geomembranes—Identification and Performance Testing
Edited by A.L.Rollin and J.M.Rigo
Ground Improvement
Edited by M.P.Mosely
Ground Subsidence
A.C.Waltham
Introducing Groundwater
M.Price
Pile Design and Construction Practice
M.J.Tomlinson
Rock Mechanics for Underground Mining
B.H.G.Brady and E.T.Brown
Soil Mechanics
R.F.Craig
Structural Foundations Manual for Low-Rise Buildings
M.F.Atkinson
The Stability of Slopes
E.N.Bromhead
Underpinning and Retention
Edited by S.Thorburn and G.S.Littlejohn
For information on these and other titles please contact:
The Promotion Department, E & FN Spon, 2–6 Boundary Row London SE1
8HN. Telephone: 071–865 0066
Engineering Treatment of
Soils
F.G.Bell
Department of Geology and Applied Geology,
University of Natal, Durban
Published by E & FN Spon in 1993.
E & FN Spon is an imprint of the Taylor & Francis Group
This edition published in the Taylor & Francis e-Library, 2005.
“To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of
thousands of eBooks please go to www.eBookstore.tandf.co.uk.”
© 1993 F.G.Bell
ISBN 0-203-47418-X Master e-book ISBN
ISBN 0-203-78242-9 (Adobe eReader Format)

ISBN 0 419 17750 7 (Print Edition)
Apart from any fair dealing for the purposes of research or private study, or criticism or
review, as permitted under the UK Copyright Designs and Patents Act, 1988, this
publication may not be reproduced, stored, or transmitted, in any form or by any means,
without the prior permission in writing of the publishers, or in the case of repromedia
reproduction only in accordance with the terms of the licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries concerning
reproduction outside the terms stated here should be sent to the publishers at the UK
address printed on this page.
The publisher makes no representation, express or implied, with regard to the accuracy of
the information contained in this book and cannot accept any legal responsibility or liability
for any errors or omissions that may be made.
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication data
Bell, F.G. (Frederic Gladstone)
Engineering treatment of soils/F.G.Bell.—1st ed.
p. cm.
Includes bibliomediaal references and index.
ISBN 0-419-17750-7 (alk. paper)
1. Soil mechanics. I. Title.
TA710.B425 1993
621.1′5136–dc20 92–40044
CIP
Contents
Preface ix
1 Introduction 1
1.1 Site investigation 2
1.2 Soil classification 3
1.3 Soil types 12
1.3.1 Coarse-grained soils 12
1.3.2 Silts and loess 13
1.3.3 Clay deposits 15
1.3.4 Tropical soils 18
1.3.5 Tills and other glacial deposits 21
1.3.6 Organic soils: peat 24
1.3.7 Fills 25
1.3.8 Coarse colliery discard 27
2 Exclusion techniques 29
2.1 Sheet piles 31
2.2 Contiguous bored piles and secant piles 33
2.3 Slurry trenches 35
2.4 Diaphragm walls 40
2.5 The character of bentonite slurry 48
2.6 Compressed air 51
3 Ground freezing 52
3.1 Introduction 52
3.2 Methods of ground freezing 54
vi
3.3 Hydrogeology and the nature of frozen soils 59

3.4 Strength and behaviour of frozen ground 61
4 Drainage techniques 67
4.1 Filter drains 67
4.2 Drainage of slopes 73
4.3 Sand drains, sandwicks and band drains 75
4.4 Lime columns 86
5 Groundwater lowering 87
5.1 Sumps 87
5.2 Wellpoint systems 88
5.3 Ejector or eductor systems 96
5.4 Bored wells 99
5.5 Settlement and groundwater abstraction 102
5.6 Electro-osmosis and electrochemical stabilization 102
6 Compaction and consolidation techniques 108
6.1 Mechanical compaction of soil 108
6.1.1 Introduction 108
6.1.2 Types of compaction equipment 110
6.1.3 Compaction and soil types 113
6.2 Precompression 118
6.3 Compaction piles 125
6.4 Vibrocompaction 127
6.4.1 Vibroflotation in granular soils 127
6.4.2 The terra-probe 132
6.4.3 Vibroreplacement 133
6.4.4 Concrete columns 137
6.4.5 Vibrodisplacement 139
6.5 Dynamic compaction 140
6.6 Explosive compaction 145
vii
7 Soil reinforcement and soil anchors 151

7.1 Reinforced earth 151
7.1.1 Introduction 151
7.1.2 Load transfer mechanism and strength development 158
7.1.3 Soil types and reinforced earth 160
7.1.4 Anchored earth 161
7.2 In-situ reinforcement techniques 162
7.2.1 Soil nailing 164
7.2.2 Reticulated micropiles 167
7.2.3 Soil dowels 169
7.3 Soil anchors 170
8 Geosynthetic materials 174
8.1 Types of geosynthetic materials 174
8.2 Properties of geosynthetic materials 177
8.2.1 Physical properties 178
8.2.2 Mechanical properties 178
8.2.3 Environmental properties 182
8.2.4 Properties characteristic of geogrids 184
8.3 Functions of geosynthetics 185
8.4 Uses of geosynthetic materials 187
9 Grouts and grouting 197
9.1 Introduction 197
9.2 Properties of grouts 199
9.3 Penetration of grouts 209
9.4 Types of grout 211
9.4.1 Cement grouts 211
9.4.2 Cement-clay grouts 215
9.4.3 Clay grouts 216
9.4.4 Clay-chemical grouts 217
viii
9.4.5 Chemical grouts 218

9.4.6 Bitumen emulsion grouts 222
9.5 Grouting 223
9.5.1 Permeation grouting 224
9.5.2 Claquage grouting 225
9.5.3 Compaction grouting 226
9.5.4 Jet grouting 227
9.5.5 Some other aspects of grouting 231
10 Soil stabilization 235
10.1 Cement stabilization 235
10.1.1 Types of soil 236
10.1.2 Mixing and compaction 238
10.1.3 Properties of cement stabilized soils 238
10.1.4 Applications of soil-cement 245
10.1.5 Plastic soil-cement 245
10.2 Lime stabilization 246
10.2.1 Properties of lime-stabilized soils 248
10.2.2 Uses of lime stabilization 251
10.3 Other materials used for soil stabilization 257
10.4 Thermal stabilization 260
Appendix: Summary of soil-treatment methods 263

References 271
Further reading 285
Index 290
Preface
The engineering treatment of soils involves improving their geotechnical

character for construction purposes. Problems regarding construction in soils
arise from their lack of strength which manifests itself in their deformation,
which beneath foundations takes the form of settlement, or in some exceptional
circumstances gives rise to ground failure. Problems are also associated with
water in soil. Hence soil treatment is primarily concerned with enhancing its
strength or reducing or excluding the water content. The techniques used to
improve soil conditions may be temporary or permanent. For example,
groundwater lowering techniques and freezing are generally temporary in nature
whilst grouting is permanent. Also some techniques have a long history whilst
others have evolved during the last 25 years, notably some of the methods of soil
reinforcement, and dynamic compaction. Moreover new techniques are
continually being developed.
Some of the techniques used to treat soils are highly specialized and as such may
not be very familiar to many of those involved with construction, notably civil
and structural engineers and engineering geologists, and to a lesser extent
builders and architects. Accordingly this text attempts to provide those in, or
intending to enter, the construction industry with a basic appreciation of the
various techniques concerned with soil treatment, under which circumstances a
particular technique will be used, and in which soil types a given method can be
used successfully.
Obviously the depth to which the subject can be dealt with is limited by the
space available (in fact the book is about one third larger than the publishers
initially agreed to). Consequently no attempt has been made to cover detailed
design methods of the various techniques mentioned. Some design methods, such
as those used in soil reinforcement are dealt with in specialist textbooks, others
such as some of those used in ground freezing may have been developed by, and
so are peculiar to, the specialist contractors concerned. However, for those who
are interested in pursuing the subject to greater depth there is a list of suggested
further readings, together with an extensive list of references. The reader will see
from the references that some journals and proceedings such as Ground
Engineering; Proceedings of the American Society of Civil Engineers—Journal
x
of the Geotechnical Engineering Division; and the Proceedings of the

International Conferences on Soil Mechanics and Foundation Engineering are
mentioned frequently. These, together with specialist journals or conference
proceedings such as the Journal of Geotextiles and Geomembranes or the
international conferences on Ground Freezing and Geotextiles, provide
particularly worthwhile sources to keep one up to date.
F.G.Bell, 1992
1
Introduction
Most civil engineering operations are carried out in soil and, obviously, poor soil
conditions will be encountered on some construction sites. If such soil cannot be
removed, then its engineering behaviour can often be enhanced by some method
of ground treatment (Glossop, 1968). Poor soil conditions usually are attributable
to an excess of ground water or a lack of strength, and associated deformability.
Treatment methods are therefore aimed at preventing ingress of groundwater to
or removing it from the site in question on the one hand or improving soil
strength on the other.
Groundwater flow becomes more significant as the permeability of the soil
increases. The fact that groundwater flow is much slower in silts and clays than
in sands and gravels does not mean that there are no problems associated with
groundwater in the two former soil types. For example, quick conditions and
piping are associated with silts and ground heave with expanding clay soils.
Conversely, inundation with water can lead to some loess soils collapsing. Other
problems associated with groundwater are dissolution of minerals such as
gypsum in gypsiferous sands and the hydration of others, for example, anhydrite.
Yet other minerals break down rapidly in the presence of groundwater, like
pyrite; the resulting sulphate ions, when carried in solution, can attack concrete
foundations.
Soils with low strength are also highly deformable. Lack of strength leads to
soil failing if it is overloaded. However, this is not a frequent occurrence in civil
engineering construction. Much more important is soil deformation which, in
terms of structures erected at the ground surface, gives rise to settlement. Some of
the most problematic soils include: peat and organic soils; quick clays, residual
montmorillonitic clays and varved clays, which may be sensitive to extra-
sensitive; and loosely packed saturated alluvial, estuarine or marine sands, silts
and muds.
Soil treatment techniques may be either temporary or permanent. For
example, the use of a wellpoint system for dewatering can be regarded as a
temporary technique, as can freezing. On the other hand, grouting is a permanent
method of ground treatment. The type of technique chosen depends on the nature
2 ENGINEERING TREATMENT OF SOILS
of the problem and the type of soil conditions. Cost is obviously a factor that
enters into the equation.
Soil is one of the most abundant and cheapest of construction materials. Even
so its use can be greatly extended by enhancing its engineering performance, for
example, by the addition of cementitious material or by incorporation of
reinforcing elements. The concept of soil reinforcement was developed over 20
years ago and is now commonly employed in soil structures, notably in
embankments and retaining walls. In this context, the use of geosynthetics has
expanded enormously but they can be used for other purposes such as filtration
and separation.
1.1
SITE INVESTIGATION
Any important ground improvement works must be preceded by a site
investigation to establish the type and succession of soils that occur at the site
concerned. The methods of investigation necessitated by a programme of ground
improvement involve those methods which are generally used in soil
investigation (Clayton et al., 1982; Weltman and Head, 1983). Obviously an
evaluation and selection of the most suitable improvement technique can only be
made after a thorough picture of the ground conditions is established. Some of
the investigation techniques used prior to the ground improvement operation
should be able to be used during and after the improvement works so that the
effectiveness of the work can be assessed.
In some cases an investigation may need to be more extensive when the same
technique is being used, but for different purposes. For example, investigations
prior to soils being grouted will probably need to be more extensive when this
method is being used to form a cutoff curtain for a dam than when it is used to
enhance the ground for foundation purposes.
In addition to establishing the geology of the site, the investigation might also
have to unravel the history of the site. This is especially the case in urban areas
where the former use of the site may be important. For example, there may be
obstructions, voids, waste materials etc., present which may adversely affect the
use of certain improvement measures. Full-scale testing on site should be carried
out if the applicability of the method in particular conditions cannot be
confirmed by routine tests.
Samples are required for conventional laboratory testing to ascertain the
properties of the soils concerned. This not only aids the selection of the
treatment process but also is required for the design of the ground improvement
programme. For instance, data on consolidation are required when the soil is to be
improved by preloading.
Ground improvement works frequently need monitoring while they are being
carried out. Again taking preloading, the amount of settlement needs to be
INTRODUCTION 3
monitored during the operation. Similarly, pore pressures in clayey soils need to
be monitored between tamping runs during dynamic compaction.
Then when the ground investigation programme has been completed, the
effectiveness of the treatment may need to be assessed. This is not the case if the
method of ground improvement employed is a temporary method. Various
methods can be employed from taking samples and testing to see if the strength
has increased or if the permeability has been reduced, to carrying out some field
test (e.g. permeability testing after grouting; pressuremeter tests for dynamic
compaction; standard or cone penetration tests after vibroflotation). Seismic
methods have also been used to assess ground improvement including crosshole
seismic, crosshole radar and interborehole acoustic emission. A survey of some
of these methods and recent developments has been provided by Rathmeyer and
Saari (1983).
1.2
SOIL CLASSIFICATION
Casagrande (1948) advanced one of the first comprehensive engineering
classifications of soil. In the Casagrande system the coarse-grained soils are
distinguished from the fine on a basis of particle size. Gravels and sands are the
two principal types of coarse-grained soils and in this classification both are
subdivided into five subgroups on a basis of grading (Table 1.1). Each of the
main soil types and subgroups are given a letter, a pair of which are combined in
the group symbol, the former being the prefix, the latter the suffix. Fine-grained
soils are subdivided on a basis of their plasticity (Table 1.1). Subsequently the
Unified Soil Classification (Table 1.2) was developed from the Casagrande
system.
The British Soil Classification for engineering purposes (Anon., 1981(b)) also
uses particle size as a fundamental parameter. Classification can be made either
by rapid assessment in the field or by full laboratory procedure (Tables 1.3 and
1.4 respectively). Boulders (B; over 200 mm), cobbles (Cb; 60–200 mm), gravels
(G; 2–60 mm), sands (S; 0.06–2 mm), silts (M; 0.002–0.06 mm) and clays (C;
less than 0.002 mm) are distinguished as individual groups. Mixed soil types are
given in Figure 1.1.
These major soil groups are again divided into subgroups on a basis of
Figure 1.1 (a) Grading triangle for soil classification (material finer than 60 mm); (b)
Fuller description of gravels and sands. (Reproduced by permission of the Chief
Executive, Transport Research Laboratory, Crowthorne, Berkshire)
INTRODUCTION 5
Table 1.1 Symbols used in the Casagrande soil classification

Main soil type Prefix
Coarse-grained soils Gravel G
Sand S
Fine-grained soils Silt M
Clay C
Organic silts and clays O
Fibrous soils Peat Pt
Subdivisions Suffix
For coarse-grained soils Well graded, with little or no fines W
Well graded with suitable clay binder C
Uniformly graded with little or no fines U
Poorly graded with little or no fines P
Poorly graded with appreciable fines or well graded F
with excess fines
For fine-grained soils Low compressibility (plasticity) L
Medium compressibility (plasticity) I
High compressibility (plasticity) H
grading in the case of cohesionless soils, and on a basis of plasticity in the case
of fine material. Granular soils are described as well graded (W) or poorly
graded (P). Two further types of poorly graded granular soils are recognized,
namely, uniformly graded (Pu) and gap-graded (Pg). Silts and clays are generally
subdivided according to their liquid limits (LL) into low (under 35%),
intermediate (35–50%), high (50–70%), very high (70–90%) and extremely high
(over 90%) subgroups. As in the Casagrande classification, each subgroup is
given a combined symbol in which the letter describing the predominant size
fraction is written first (e.g. GW=well-graded gravels; CH=clay with high liquid
limit; see Table 1.4).
Any group may be referred to as organic if it contains a significant proportion
of organic matter, in which case the letter O is suffixed to the group symbol (e.g.
CVSO=organic clay of very high liquid limit with sand). The symbol Pt is given
to peat.
In many soil classifications boulders and cobbles are removed before an
attempt is made at classification; for example, their proportions are recorded
separately in the British Soil Classification. Their presence should be recorded in
the soil description, a plus sign being used in symbols for soil mixtures, for
example, G+Cb for gravel with cobbles. The British Soil Classification groups
very coarse deposits as follows:
1. Boulders Over half of the very coarse material is of boulder size (over 200
mm); may be described as cobbly boulders if cobbles are an important
second constituent in the very coarse fraction.
INTRODUCTION 7
INTRODUCTION 9
Anon. 1981(b))
INTRODUCTION 11
Table 1.5 Description of material/soil mixtures

Term Composition
Boulders (or Cobbles) with a little Up to 5% finer material

finer material*
Boulders (or Cobbles) with some 5–20% finer material
finer material*
Boulders (or Cobbles) with much 20–50% finer material
finer material*
Finer material* with many 50–20% boulders (or cobbles)
Boulders (or Cobbles)
Finer material* with some 20–5% boulders (or cobbles)
Finer material* with occasional Up to 5% boulders (or cobbles)
* Give the name of the finer material (in parentheses when it is is the minor
constituent), e.g. cobbly boulders with some finer material (sand
with some fines).
2. Cobbles Over half of the very coarse material is of cobble size (200–60
mm); may be described as bouldery cobbles if boulders are an important
second constituent in the very coarse fraction.
Mixtures of very coarse material and soil can be described by combining the
terms for the very coarse constituent and the soil constituent, as shown in
Table 1.5.
1.3
SOIL TYPES
1.3.1
Coarse-grained soils
Size and sorting have a significant influence on the engineering behaviour of
granular soils. For example, the void ratios of well-sorted and perfectly
cohesionless aggregates of equidimensional grains can range between values of
about 0.35 and 1.00. If the void ratio is more than unity the microstructure is
collapsible or metastable. Generally speaking the larger the particles, the higher
the strength, and deposits consisting of a mixture of different-sized particles are
usually stronger than those that are uniformly graded (Table 1.6).
Densely packed sands are almost incompressible, whereas loosely packed
deposits, located above the water table, are relatively compressible but otherwise
stable. If the relative density of a sand varies erratically, this can give rise to
differential settlement. Greater settlement is likely to be experienced in granular
soils when foundation level is below the water table than when above.
Settlement is relatively rapid in granular soils.
INTRODUCTION 13
1.3.2
Silts and loess
The grains in a deposit of silt are often rounded with smooth outlines. This
influences their degree of packing. The latter, however, is more dependent on the
grain size distribution within a silt deposit, uniformly sorted deposits not being
able to achieve such close packing as those in which there is a range of grain size.
This, in turn, influences the values of void ratio (0.35–0.85) as well as the bulk
and dry densities (Table 1.6).
Dilatancy is characteristic of fine sands and silts. The environment is all-
important for the development of dilatancy since conditions must be such that
expansion can take place.
Consolidation of silt is influenced by grain size, particularly the size of the
clay fraction, porosity and natural moisture content. Primary consolidation may
account for over 75% of total consolidation. In addition, construction settlement
may continue for several months after completion because the rate at which
water can drain from the voids under the influence of applied stress is slow.
Loess owes its engineering characteristics largely to the way in which it was
deposited since this commonly has given it a metastable structure, in that initially
the particles were loosely packed. The porosity of the structure is enhanced by
the presence of fossil root-holes. The latter are lined with carbonate cement,
which helps bind the grains together. This means that the initial, loosely packed
structure is preserved and the carbonate cement provides some of the bonding
strength of loess. However, the chief binder is usually the clay matrix. On
wetting, the clay bond in many loess soils becomes soft, which can lead to the
collapse of the metastable structure. The breakdown of the soil structure can
occur under its own weight.
Loess deposits generally consist of 50–90% particles of silt size. In fact sandy,
silty and clayey loess can be distinguished. The undisturbed densities of loess
may range from around 1.2–1.36 t/m3. If wetted (or reworked), the density of
collapsible loess increases, sometimes to as high as 1.6 t/m3. The liquid limit of
Table 1.6 Some typical properties of gravels, sands and silts
Gravels Sands Silts
Relative density 2.5–2.8 2.6–2.7 2.64–2.66
Bulk density (t/m3) 1.45–2.3 1.4–2.15 1.82–2.15
Gravels Sands Silts

Dry density (t/m3) 1.4–2.1 1.35–1.9 1.45–1.95
Porosity (%) 20–50 23–35 –
Void ratio – – 0.35–0.85
Liquid limit (%) – – 24–35
Plastic limit (%) – – 14–25
Coefficient of consolidation (m2/yr) – – 12.2
Effective cohesion (kPa) – – 75
Shear strength (kPa) 200–600 100–300 –
Angle of friction (deg) 35–45 32–42 32–36
Figure 1.2 Particle size distribution and plasticity of Missouri river basin loess (after
Clevenger, 1958, with permission of ASCE)
loess averages about 30% (exceptionally liquid limits as high as 45% have been
recorded), and their plasticity index ranges from about 4 to 9%, but averages 6%
(Figure 1.2).
Normally loess possesses a high shearing resistance and can carry high
loadings without significant settlement when natural moisture contents are low.
For instance, natural moisture contents of undisturbed loess are generally around
INTRODUCTION 15
10% and the supporting capacity of loess at this moisture content is high.
However, the density of loess is the most important factor controlling its shear
strength and settlement. On wetting, large settlements and low shearing
resistance are encountered when the density of loess is below 1.30 t/m3, whereas
if the density exceeds 1.45 t/m3 settlement is small and shearing resistance is
fairly high.
Loess deposits are better drained (their permeability ranges from 10–5 to 10–7
m/s) than are true silts because of the fossil root-holes. Their permeability is
appreciably higher in the vertical than in the horizontal direction.
1.3.3
Clay deposits
The principal minerals in a deposit of clay tend to influence its engineering
behaviour. For example, the plasticity of a clay soil is influenced by the amount
of its clay fraction and the type of clay minerals present since clay minerals
greatly influence the amount of attracted water held in a soil. The undrained shear
strength is related to the amount and type of clay minerals present in a clay
deposit together with the presence of cementing agents. In particular, strength is
reduced with increasing content of mixed-layer clay and montmorillonite in the
clay fraction. The increasing presence of cementing agents, especially calcite,
enhances the strength of the clay.
Geological age also has an influence on the engineering behaviour of a clay
deposit. The porosity, water content and plasticity normally decrease in value
with increasing depth, whereas the strength and elastic modulus increase.
The engineering performance of clay deposits is also affected by the total
moisture content and by the energy with which this moisture is held. For
instance, the moisture content influences their consistency and strength, and the
energy with which moisture is held influences their volume change
characteristics.
One of the most notable characteristics of clays from the engineering point of
view is their susceptibility to slow volume changes that can occur independent of
loading due to swelling or shrinkage. Differences in the period and magnitude of
precipitation and evapotransportation are the major factors influencing the swell-
shrink response of a clay beneath a structure. Generally kaolinite has the smallest
swelling capacity of the clay minerals. Illite may swell by up to 15% but
intermixed illite and montmorillonite may swell some 60–100%. Swelling in Ca
montmorillonite is very much less than in the Na variety; it ranges from about 50
to 100%. Swelling in Na montmorillonite occasionally can amount to 2000% of
the original volume. One of the most widely used soil properties to predict swell
potential is the activity of a clay (Figure 1.3).
Volume changes in clays also occur as a result of loading and unloading which
bring about consolidation and heave, respectively. When a load is applied to a
clay soil its volume is reduced, this being due principally to a reduction in the
Figure 1.3 Estimation of the degree of expansiveness of a clay soil (after Williams and
Donaldson, 1980)
void ratio. If such a soil is saturated, then the load is initially carried by the pore
water which causes a pressure, the hydrostatic excess pressure, to develop. The
excess pressure of the pore water is dissipated at a rate which depends upon the
permeability of the soil mass, and the load is eventually transferred to the soil
structure. Primary consolidation is brought about by a reduction in the void ratio.
Further consolidation may occur due to a rearrangement of the soil particles.
This secondary compression is usually much less significant. In clay soils,
because of their low permeability, the rate of consolidation is slow. The
compressibility of a clay is related to its geological history, that is, to whether it
is normally consolidated or overconsolidated.
The heave potential arising from stress release depends upon the nature of the
diagenetic bonds within the soil. For example, when an excavation is made in a
clay with weak diagenetic bonds, elastic rebound causes immediate dissipation
of some stored strain energy in the soil, this being manifested in a certain amount
of heave.
An overconsolidated clay is considerably stronger at a given confining
pressure than a normally consolidated clay and tends to dilate during shear,
whereas a normally consolidated clay consolidates. Hence, when an
overconsolidated clay is sheared under undrained conditions negative pore water
INTRODUCTION 17
Figure 1.4 Peak strength and residual strength of normally (N-C) and overconsolidated
(O-C) clays (after Skempton, 1964)
pressures are induced, the effective strength is increased, and the undrained
strength is much higher than the drained strength—the exact opposite to a
normally consolidated clay. When the negative pore-water pressure gradually
dissipates the strength falls as much as 60 or 80% to the drained strength.
Skempton (1964) observed that when clay is strained it develops an increasing
resistance (strength), but that under a given effective pressure the resistance
offered is limited, the maximum value corresponding to the peak strength. If
testing is continued beyond the peak strength, then, as displacement increases,
the resistance decreases, again to a limiting value which is termed the residual
strength. In moving from peak to residual strength, cohesion falls to almost, or
actually, zero and the angle of shearing resistance is reduced to a few degrees (it
may be as much as 10° in some clays). Under a given effective pressure, the
residual strength of a clay is virtually the same whether it is normally
consolidated or overconsolidated (Figure 1.4). Furthermore, the value of residual
shear strength ( ’r) decreases as the amount of clay fraction increases in a
deposit. Not only is the proportion of detrital minerals important but so is that of
the diagenetic minerals. The latter influence the degree of induration of a deposit
of clay and the value of ’r can fall significantly as the ratio of clay minerals to
detrital and diagenetic minerals increases.
The shear strength of an undisturbed clay is frequently found to be greater than
that obtained when it is remoulded and tested under the same conditions and at
the same water content. The ratio of the undisturbed to the remoulded strength at
the same moisture content is termed the sensitivity of a clay. Clays with high
sensitivity values have little or no strength after being disturbed. Sensitive clays
generally possess high moisture contents, frequently with liquidity indices well
in excess of unity. A sharp increase in moisture content may cause a great
increase in sensitivity, sometimes with disastrous results. Heavily
overconsolidated clays are insensitive.
Fissures in clays play an extremely important role in their failure mechanism
and are characteristic of overconsolidated clays. For example, the strength along
fissures in the clay is only slightly higher than the residual strength of the intact
clay. Hence, the upper limit of the strength of fissured clay is represented by its
intact strength while the lower limit corresponds to the strength along the
fissures. The operational strength is, however, often significantly higher than the
fissure strength. The ingress of water into fissures means that the pore-water
pressure in the clay concerned increases, which in turn means that its strength is
reduced.
Fissures in overconsolidated clays can have practical consequences in that the
strength of the clays can be significantly reduced. For example, Skempton et al.
(1969) summarized the shear strength parameters of London Clay (which is a
heavily overconsolidated fissured clay) in terms of effective stress as follows:
1. Peak strength of intact clay: c’=31 kPa, '=20;

2. ‘Peak’ strength on fissure and joint surfaces: c'=6.9 kPa, '=18.5;
3. Residual strength of intact clay: c’r=1.4 kPa, ’r=16.
The greatest variation in the engineering properties of clays can be attributed to

the degree of weathering they have undergone. Ultimately weathering, through
the destruction of interparticle bonds, leads to a clay deposit reverting to a normally
consolidated, sensibly remoulded condition. Higher moisture contents are found
in the more weathered clay. This progressive degrading and softening is also
accompanied by reductions in strength and deformation moduli with a general
increase in plasticity.
1.3.4
Tropical soils
Ferruginous and aluminous clay soils are frequent products of weathering in
tropical latitudes (Anon., 1990(a)). They are characterized by the presence of
iron and aluminium oxides and hydroxides. Laterite is a residual ferruginous
clay-like deposit which generally occurs below a hardened ferruginous crust or
hardpan. During drier periods the water table is lowered. The small amount of
iron that has been mobilized in the ferrous state by the groundwater is then
oxidized, forming haematite or, if hydrated, goethite. The movement of the water
table leads to the gradual accumulation of iron oxides at a given horizon in
the soil profile. A cemented layer of laterite is formed which may be a
continuous or honeycombed mass, or nodules may be formed, as in laterite gravel.
Laterite hardens on exposure to air. Hardening may be due to a change in the
hydration of iron and aluminium oxides. It commonly contains all size fractions
from clay to gravel and sometimes even larger material (Figure 1.5). Values of
common properties of laterite are given in Table 1.7. Such soils are of low to
medium plasticity. The strength of laterite may decrease with increasing depth
beneath the hardened crust
INTRODUCTION 19
Figure 1.5 Grading curves of laterites (after Madu, 1977)
Table 1.7 Some common properties of laterites

Moisture content (%) 10–49
Liquid limit (%) 33–90
Plastic limit (%) 13–31
Clay fraction 15–45
Dry unit weight (kN/m3) 15.2–17.3
Cohesion, cu(kPa) 466–782
Angle of internal friction, u (°) 28–35
Unconfined compressive strength (kPa) 220–625
Compression index 0.0186
Coefficient of consolidation (m2/year) 262–599*
Young’s modulus (kPa) 5.63×104
* For a pressure of 215 kPa
(where present). The latter has a low compressibility and settlement, therefore,
is usually negligible.
Red earths or latosols are residual ferruginous soils in which oxidation readily
occurs. Most of them appear to have been derived from the first cycle of
weathering of the parent material. They differ from laterite in that they behave as
a clay and do not possess strong concretions. They do, however, grade into
laterite.
Black clays are typically developed on poorly drained soils in regions with
well-defined wet and dry seasons, where the annual rainfall is not less than 1250
mm. Generally the clay fraction in these soils exceeds 50%, silty material
varying between 20 and 40%, and sand forming the remainder. The organic
content is usually less than 2%. The liquid limits of black clays may range
between 50 and 100%, with plasticity indices of between 25 and 70%. The
shrinkage limit is frequently around 10–12%. Montmorillonite is commonly
present in the clay fraction and is the chief factor determining the behaviour of
these clays. For instance, they undergo appreciable volume changes on wetting
and drying due to the montmorillonite content. These volume changes, however,
tend to be confined to an upper critical zone of the soil, which is frequently less
than 2.5 m thick. Below this the moisture content remains more or less the same,
for instance, around 25%.
Dispersive clay soils deflocculate in the presence of relatively pure water to
form colloidal suspensions and therefore are highly susceptible to erosion and
piping. Piping is initiated by dispersion of clay particles along desiccation
cracks, fissures and root-holes. There is no threshold velocity for dispersive clay,
the colloidal clay particles go into suspension even in quiet water. Hence
retrogressive erosion can occur at very low pore-water flow velocities. Such soils
contain a higher content of dissolved sodium in their pore water than ordinary
soils (up to 12%, with pH values varying between 6 and 8). There are no
significant differences in the clay contents of dispersive and non-dispersive soils,
except that soils with less than 10% clay particles may not have enough colloids
to support dispersive piping. Potentially dispersive soils frequently contain a
moderate to high content of clay.
Dispersive erosion depends on the mineralogy and chemistry of the soil on the
one hand, and the dissolved salts in the pore and eroding water on the other. The
presence of exchangeable sodium is the main chemical factor contributing to
dispersive clay behaviour. This is expressed in terms of the exchangeable sodium
percentage (ESP= exchangeable sodium÷cation exchange capacity, %) where the
units are given in meq./100 g of dry soil. Above threshold values of ESP of 10%,
soils that have their free salts leached by seepage of relatively pure water are
prone to dispersion. Soils with ESP values above 15% are highly dispersive.
High ESP values and piping potential generally exist in soils in which the clay
fraction is composed largely of smectitic and other 2:1 clays. Some illitic soils
are highly dispersive. High values of ESP and high dispersibility are rare in clays
composed largely of kaolinites.
The main property of clay soils governing their susceptibility to dispersion is
the total content of dissolved salts in the water. The lower the content of
dissolved salts in the water, the greater the susceptibility of sodium-saturated
clay to dispersion. There is a threshold value for total cation concentration
(TCC) in the pore water (for a given ESP) above which the soil remains
flocculated. Figure 1.6 shows the zones in which a soil of a given ESP can exist
in either a dispersed or flocculated state, depending on the concentration of the
INTRODUCTION 21
Figure 1.6 The influence of free salts on dispersion and flocculation as influenced by the
colloidal composition of soils. SAR=Na/[(Ca+Mg)/2]
salts in the pore water. A number of special tests are used to identify dispersive
soils. They have been reviewed by Sherard et al. (1976) and Gerber and Harmse
(1987).
Calcareous silty clays are important types of soil in arid and semi-arid areas.
These silty clays are light to dark brown in colour. They normally are formed by
deposition in saline or lime-rich waters. These soils possess a stiff to hard
desiccated clay crust, referred to as duricrust, which may be up to 2 m thick, and
which overlies moist soft silty clay.
In arid and semi-arid regions the evaporation of moisture from the surface of
the soil may lead to the precipitation of salts in the upper layers. The most
commonly precipitated material is calcium carbonate. These caliche deposits are
referred to as calcrete. The development of calcrete is inhibited beyond a certain
aridity since the low precipitation is unable to dissolve and drain calcium
carbonate towards the water table. Consequently in arid climates gypcrete may
take the place of calcrete.
The hardened calcrete crust may contain nodules of limestone or be more or less
completely cemented (this cement may, of course, have been subjected to
differential leaching). As the carbonate content increases it first occurs as
scattered concentrations of flaky habit, then as hard concretions. Once it exceeds
60%, the concentration becomes continuous. The calcium carbonate in calcrete
profiles decreases from top to base, as generally does the hardness.
1.3.5
Tills and other glacial deposits
Till is usually regarded as being synonymous with boulder clay. It is deposited
directly by ice while stratified drift is deposited in melt waters associated with
glaciers. The character of till deposits varies appreciably and depends mainly on
the lithology of the material from which it was derived. The underlying bedrock
material usually contributes up to about 80% of basal or lodgement tills.

Lodgement till is plastered on to the ground beneath a moving glacier in small
increments as the basal ice melts. Because of the overlying weight of ice such
deposits are overconsolidated.
The proportion of silt and clay size material is relatively high in lodgement till
(e.g. the clay fraction varies from 15 to 40%). Lodgement till is commonly stiff,
dense and relatively incompressible. Hence it is practically impermeable. Fissures
are frequently present in lodgement till, especially if it is clay matrix dominated.
Ablation till accumulates on the surface of the ice when englacial debris melts
out, and as the glacier decays the ablation till is slowly lowered to the ground. It
is therefore normally consolidated. Because it has not been subjected to much
abrasion, ablation till is characterized by abundant large stones that are angular,
the proportion of sand and gravel is high and clay is present only in small
amounts (usually less than 10%). Because the texture is loose, ablation till can
have an extremely low in-situ density. Since ablation till consists of the load
carried at the time of ablation it usually forms a thinner deposit than lodgement
till.
Tills are frequently gap graded, the gap generally occurring in the sand
fraction (Figure 1.7). Large, often very local, variations can occur in the grading
of till. The range in the proportions of coarse and fine fractions in tills dictates
the degree to which the properties of the fine fraction influence the properties of
the composite soil. The variation in the engineering properties of the fine soil
fraction is greater than that of the coarse fraction, and this often tends to
dominate the engineering behaviour of the till.
The consistency limits of tills are dependent upon water content, grain size
distribution and the properties of the fine-grained fraction. Generally, however,
the plasticity index is small and the liquid limit of tills decreases with increasing
grain size. Dense, heavily overconsolidated till is relatively incompressible and
when loaded undergoes very little settlement, most of which is elastic. For the
average structure such elastic compressions are too small to be of concern and
therefore can be ignored.
Opening up and softening along fissures gives rise to a rapid reduction of
undrained shear strength along the fissures. Deformation and permeability are
also controlled by the nature of the fissures. Fissures tend to be variable in
character, spacing, orientation and areal extent.
Deposits of stratified drift are often subdivided into two categories, namely,
those which develop in contact with the ice—the ice contact deposits—and those
which accumulate beyond the limits of the ice, forming in streams, lakes or seas
—the proglacial deposits.
Outwash fans range in particle size from coarse sands to boulders. When they
are first deposited their porosity may be anything from 25 to 50% and they tend
to be very permeable. The finer silt-clay fraction is transported further
downstream. Kames, kame terraces and eskers usually consist of sands and
gravels.
INTRODUCTION 23
Figure 1.7 Typical gradings of some Scottish morainic soils (after McGown, 1971)
The most familiar proglacial deposits are varved clays. The thickness of the
individual varve is frequently less than 2 mm, although much thicker layers have
been noted in a few deposits. Generally the coarser layer is of silt size and the
finer of clay size. Varved clays tend to be normally consolidated or lightly
overconsolidated.
The range of liquid limits for varved clays tends to vary between 30 and 80%,
while that of plastic limit often varies between 15 and 30%. Hence they are
inorganic silty clays of medium to high plasticity or compressibility. In some
varved clays the natural moisture content is near the liquid limit so they are
medium-sensitive clays. The effective stress parameters of apparent cohesion and
angle of shearing resistance frequently range from 5 to 19.5 kPa and 22 to 25°
respectively.
The material of which quick clays are composed is predominantly smaller than
0.002 mm but many deposits seem to be very poor in clay minerals, containing a
high proportion of ground-down, fine quartz. Particles, whether aggregations or
individual minerals, are rarely in direct contact, being linked generally by
bridges of fine particles.
Quick clays generally exhibit little plasticity, their plasticity index generally
varying between 8 and 12%. Their liquidity index normally exceeds 1, and their
liquid limit is often less than 40%. The most extraordinary property possessed by
quick clays is their very high sensitivity.
1.3.6
Organic soils: peat
Peat is an accumulation of partially decomposed and disintegrated plant remains
which have been fossilized under conditions of incomplete aeration and high
water content. Physico-chemical and biochemical processes cause this organic
material to remain in a state of preservation over a long period of time.
Macroscopically, peaty material can be divided into three basic groups,
namely, amorphous granular, coarse fibrous and fine fibrous peat. The
amorphous granular peats have a high colloidal fraction, holding most of their
water in an adsorbed rather than free state. In the other two types the peat is
composed of fibres, these usually being woody. In the coarse variety a mesh of
second-order size exists within the interstices of the first-order network, while in
fine fibrous peat the interstices are very small and contain colloidal matter.
The mineral material in peat is usually quartz sand and silt. In many deposits
the mineral content increases with depth. The amount of mineral content
influences the engineering properties of peat.
The void ratio of peat ranges from about 9 for dense amorphous granular peat,
up to 25 for fibrous types with high contents of sphagnum. It tends to decrease
with depth within a peat deposit. Such high void ratios give rise to a
phenomenally high water content and most of the peculiarities of peat are
attributable to its moisture content. This varies according to the type of peat; it may
be as low as 500% while values exceeding 3000% occur. The amount of
shrinkage that can occur in peat generally ranges between 10 and 75% of the
original volume, and it can involve reductions in void ratio from over 12 down to
about 2.
Dry densities of drained peat fall within the range 65–120 kg/m3. The dry
density is influenced by the mineral content, and higher values than those
mentioned can be obtained when peats possess high mineral residues.
Apart from its moisture content and dry density, the shear strength of a peat
deposit appears to be influenced by its degree of humidification and its mineral
content. As both these factors increase, so does the shear strength. Conversely,
the higher the moisture content of peat, the lower is its shear strength.
In an undrained bog the unconfined compressive strength may be negligible,
the peat possessing a consistency approximating to that of a liquid. The strength
is increased by drainage to values between 20 and 30 kPa and the modulus of
elasticity to between 100 and 140 kPa. When loaded, peat deposits undergo high
deformations but their modulus of deformation tends to increase with increasing
load. If peat is very fibrous it appears to suffer indefinite deformation without
planes of failure developing, whereas failure planes nearly always form in dense
amorphous peats.
Differential and excessive settlement is the principal problem confronting the
engineer working on a peaty soil. When a load is applied to peat, settlement occurs
because of the low lateral resistance offered by the adjacent unloaded peat.
INTRODUCTION 25
Serious shearing stresses are induced even by moderate loads. Worse still, should
the loads exceed a given minimum, then settlement may be accompanied by
creep, lateral spreading or, in extreme cases, by rotational slip and upheaval of
adjacent ground. At any given time the total settlement in peat due to loading
involves settlement with and without volume change. Settlement without volume
change is the more serious for it can give rise to the types of failure mentioned.
What is more, it does not enhance the strength of peat.
1.3.7
Fills
Because suitable sites are becoming scarce in urban areas, the construction of
buildings on fill or made-up ground has assumed a greater importance. A wide
variety of materials is used for fills, including domestic refuse, ashes, slag,
clinker, building waste, chemical waste, quarry waste and all types of soils. The
extent to which an existing fill will be suitable as a foundation depends largely
on its composition and uniformity. In the past the control exercised in placing fill
has frequently been insufficient to ensure an adequate and uniform support for
structures immediately after placement. Consequently a time interval had to be
allowed prior to building so that the material could consolidate under its own
weight. Although this may be suitable for small, lightly loaded buildings it is
unsatisfactory for more heavily loaded structures that can give rise to substantial
settlement.
The time taken for a fill to reach a sufficient degree of natural consolidation so
that it becomes suitable for a foundation depends on the nature and thickness of
the fill, the method of placing and the nature of the underlying ground, especially
the groundwater conditions. The best materials in this respect are obviously well
graded, hard and granular (Figure 1.8). By contrast, fills containing a large
proportion of fine material may take a long time to settle. Generally rock fills
will settle 2.5% of their thickness, sandy fills about 5% and cohesive material
around 10%. The rate of settlement decreases with time but in some cases it may
take 10–20 years before movements are reduced within tolerable limits for
building foundations. In coarse-grained soils the larger part of movement
generally occurs within the first two years after the construction of the fill, and
after five years settlements are usually
Figure 1.8 Observations of the settlement of various types of fill due to consolidation under
its own weight (after Meyerhof, 1951). Description of curves: 1, well-graded sand, well
compacted; 2, rockfill, medium state of compaction; 3, clay and chalk, lightly compacted;
4, sand, uncompacted; 5, clay, uncompacted; 6, mixed refuse, well compacted
Table 1.8 Municipal waste materials incorporated in fills (after Sowers, 1973)
Material Characteristics as fill
1. Garbage: food, waste Wet. Ferments and decays readily.
Compressible, weak
2. Paper, cloth Dry to damp. Decays and burns.
Compressible
3. Garden refuse Damp. Ferments, decays, burns.
Compressible
4. Plastic Dry. Decay resistant, may burn. Compressible
5. Hollow metal, e.g. drums Dry. Corrodable and crushable
6. Massive metal Dry. Slightly corrodable. Rigid
7. Rubber, e.g. tyres Dry. Resilient, burns, decay resistant.
Compressible
8. Glass Dry. Decay resistant. Crushable and
compressible
9. Demolition timber Dry. Decays and burns. Crushable
INTRODUCTION 27
Material Characteristics as fill

10. Building rubble Damp. Decay resistant. Crushable and
erodable
11. Ashes, clinker and chemical wastes Damp. Compressible, active chemically and
partially soluble
very small. The minimum time that should elapse before development takes
place on an opencast backfill should be 12 years after restoration is complete.
Frequently, poorly compacted old fills continue to settle for years due to secondary
consolidation.
Waste disposal or sanitary land fills are usually very mixed in composition
(Table 1.8) and suffer from continuing organic decomposition and physico-
chemical breakdown which may leave voids. Methane and hydrogen sulphide are
often produced in the process, and accumulations of these gases in pockets in
fills have led to explosions. The production of leachate is another problem. Some
material such as ashes and industrial wastes may contain sulphates and other
products which are potentially damaging to concrete. The density of waste
disposal fills varies from about 120 to 300 kg/m3 when tipped. After compaction
the density may exceed 600 kg/m3. Moisture contents range from 10 to 50% and
the average specific gravity of the solids from 1.7 to 2.5. Settlements are likely to
be large and irregular. The initial settlement of waste disposal fills is rapid and is
due to a reduction in the void ratio. It takes place with no build up of pore-water
pressure. Settlement continues due to a combination of secondary compression
(material disturbance) and physico-chemical and biochemical action.
1.3.8
Coarse colliery discard
There are two types of colliery discard, namely, coarse and fine. Coarse discard
consists of run-of-mine material and reflects the various rock types that are
extracted during mining operations. It contains varying amounts of coal that has
not been separated by the preparation process. Fine discard consists of either
slurry or tailings from the washery, which is pumped into lagoons. Some tips,
particularly those with relatively high coal contents, may be partly burned or
burning and this affects their composition and, therefore, their engineering
behaviour.
The majority of tip material is essentially granular. Often most of it falls
within the sand range, but significant proportions of gravel and cobble range may
also be present. Owing to breakdown, older and surface materials tend to contain
a higher proportion of fines than that occurring within a tip. The moisture
content of coarse discard increases with increasing content of fines, and
generally falls within the range 5–15%. Tip material shows a wide variation in
bulk density and may, in fact, vary within a spoil heap. Low densities are mainly
a function of low specific gravity.
As far as effective shear strength of coarse discard is concerned, ’ usually
varies from 25 to 45°. The angle of shearing resistance, and therefore the
strength, increases in spoil that has been burned. With increasing content of fine
coal, on the other hand, the angle of shearing resistance is reduced. The shear
strength of colliery spoil, and therefore its stability, is dependent upon the pore
pressures developed within it. These are likely to be developed where there is a
high proportion of fine material, which reduces the permeability below 5×10–7 m/
s.
Oxidation of pyrite within tip waste is governed by access of air. However, the
highly acidic oxidation products that result may be neutralized by alkaline
materials in the waste. The sulphate content of weathered, unburned colliery
waste is usually high enough to warrant special precautions in the design of
concrete structures which may be in contact with the discard.
Spontaneous combustion of carbonaceous material, frequently aggravated by
the oxidation of pyrite, is the most common cause of burning spoil. The problem
of combustion has sometimes to be faced when reclaiming old tips. Spontaneous
combustion may give rise to subsurface cavities in spoil heaps and burned ashes
may also cover zones which are red hot to appreciable depths. When steam
comes in contact with red-hot carbonaceous material, watergas is formed, and
when the latter is mixed with air it becomes potentially explosive. Explosions
may occur when burning spoil heaps are being reworked and a cloud of coal dust
is formed near the heat surface.
Noxious gases are emitted from burning spoil. These include carbon
monoxide, carbon dioxide, sulphur dioxide and, less frequently, hydrogen
sulphide. Carbon monoxide may be present in potentially lethal concentrations.
Sulphur gases are usually not present in high concentrations.
2
Exclusion techniques
Groundwater frequently provides one of the most difficult problems during

excavation. Not only does water make working conditions difficult—for
example, vehicles may become bogged down, muck becomes more difficult to
handle—but flow of water into an excavation can lead to erosion and failure of
the sides. Collapsed material has to be removed and the damage made good.
Instability of the floor may arise when deep excavations are made into sandy
ground. This is caused by upward seepage pressure of water, and piping and
boiling results when the upward seepage velocity is high enough to suspend the
individual grains of sand. Piping and boiling or quick conditions can also occur
in silt. On the other hand, the velocity of ground water flow in clay soils is
usually so slow as not to present erosion problems. Some of the worse conditions
are met in excavations that have to be taken beneath the water table.
Artesian conditions, in particular, can cause serious trouble in excavations
and, therefore, if such conditions are expected it is essential that both the position
of the water table and piezometric pressures should be determined before work
commences. Otherwise excavations which extend close to strata under artesian
pressure may be severely damaged due to blow-outs taking place in their floors.
Such action may also cause slopes to fail and, in fact, has led to the abandonment
of sites. Sites at which such problems are likely to be encountered should be
dealt with prior to and during excavation, by employing exclusion or dewatering
techniques.
In order to help select the type of cut-off required for groundwater control
about an excavation, a number of basic questions have to be answered. First, is a
cut-off the best method of groundwater control at the site concerned? Secondly,
if it is, then which cut-off technique is best suited to the site conditions? Thirdly,
can the cut-off be temporary—that is, remain in existence for the construction
period only—or has it to be permanent? Fourthly, can the cut-off serve any other
function?
Methods of forming such barriers include sheet piling, contiguous bored and
secant pile walls, bentonite cut-off walls, geomembrane barriers, concrete
diaphragm walls, grout curtains and panels, and ice walls (Bell and Mitchell,
1986). The economy of providing a barrier to exclude groundwater depends on
Figure 2.1 Sheet piling and the possibility of instability at the base of an excavation due
to upward seepage pressure.
the existence of an impermeable stratum beneath the excavation to form an

effective natural cut-off for the barrier. If such a stratum does not exist or if it
lies at too great a depth to be used as a cut-off, then upward seepage may occur
which, in turn, may give rise to instability at the base of the excavation
(Figure 2.1). In these circumstances the barrier will not be totally effective unless
it can be extended horizontally beneath the excavation. The only methods of
forming a horizontal barrier are by grouting or freezing.
In urban areas the presence of existing buildings and the high price of land
frequently dictate that construction be carried out in the least possible space.
Where surrounding property has to be safeguarded it will usually be more
appropriate to provide a barrier or cut-off around an excavation so as to prevent
the inflow of water while maintaining the water table outside at its normal level,
rather than adopt a dewatering technique. By keeping the water table at its
normal level, settlement problems are not as likely to arise as they would if a
dewatering technique was used. Exclusion techniques can be used with
dewatering techniques.
Perimeter exclusion walls should be capable of construction through soft, wet
ground; these walls should be economical in installation cost and should be
capable of being installed without any adverse effects on the stability of adjacent
buildings. If possible, they should eliminate the need for temporary work during
construction of the permanent works. Any movement of the barrier can adversely
affect both the structure to be incorporated within the excavation and the stability
of existing buildings immediately adjacent to it. Accordingly, measurements
EXCLUSION TECHNIQUES 31
should be taken of the constructed walls and adjacent properties to check for any
movements.
2.1
SHEET PILES
Sheet piling may be composed of steel, timber or concrete piles, each pile being
linked to the next to form a continuous wall. The integrity of sheet-pile walls
depends on the interlock between individual piles. Steel sheet piles are most
widely used (Figure 2.2) since they represent the simplest and cheapest method of
forming an impermeable barrier, especially if the piles can be extracted for re-
use. By contrast, precast reinforced concrete piles nowadays are seldom used.
Steel sheet piles are used in temporary structures such as strutted excavations
and cofferdams, as well as for permanent use, notably in retaining walls and
bulkheads. With all sections of pile or types of steel, higher stresses are
permissible when used in temporary works rather than for permanent use. Sheet
piling, when installed before excavation takes place, provides an effective cut-
off, as well as affording ground support in that the piles are used to resist horizontal
pressures developed by soil and pore water. The piling derives its stability from
the horizontal support provided by the ground into which it is driven, and from
any anchors, ties or struts located near the top of the piles.
Steel sheet piling is most effective as a cut-off when driven into an underlying
impermeable bed of clay. Should a thick deposit of sand exist below the clay, the
differential head should be monitored by a piezometer (installed outside the
piling) to ensure that the thickness of the bed of clay can offer adequate
resistance. If it cannot, then pressure relief wells must be provided.
Steel sheet piling generally is not used in ground containing numerous
boulders such as some tills, because the piles become difficult or impossible to
drive and there is a risk that they might tear. One technique which has proved to
be effective in such conditions involves excavating the boulders in a slurry
trench, backfilling with sand, and then driving the piles. If the sand is thoroughly
mixed before placement, the effectiveness of the piling as a cut-off is enhanced.
The pile section to be selected depends on height of ground to be retained, soil
and groundwater conditions, and on whether the piling is tied back or strutted.
The shape of the pile cross-section also is designed to make the wall capable of
resisting bending. The section has to be capable of being driven into the soil to
the desired depth and, in this context, the shape of the cross-section provides
stiffness.
The most critical parts of a sheet-pile wall for groundwater exclusion are the
interlocks which provide the connection between piles, thereby forming a
continuous wall. Such a wall generally is sufficiently watertight for most
practical purposes. However, if piles come out of interlock during driving this
means that the effectiveness of the cut-off is impaired. Moreover, when the sheet
piling remains interlocked, the cut-off is still of limited effectiveness until the
Figure 2.2 (a) Steel sheet piling with bracing in variable soil conditions; (b) steel sheet
piles used for forming a caisson in alluvial soils, near Kings Lynn, Norfolk
sheet is stressed, wedging the adjacent piles into tight contact. In fact leakage can
be quite high when a row of steel sheet piles, acting as a cut-off, is unstressed. It
is therefore usually necessary to seal the interlocks by using heavy grease before
installation, or by grouting.
Since a much larger amount of penetration below excavation level is required
for a cantilever wall than an anchored wall, it is imperative to ensure that the
section of piling finally calculated on bending moment considerations is large
enough to withstand the vertical stresses developed during driving. In the former
type of wall the depth of penetration should be at least equal to the height of the
sheet piling above excavation level.
When steel sheet piling extends into undisturbed soil, even though it goes
below the water table, corrosion tends to be very slight. Hence it is generally
unnecessary to apply a protective coating to the piles. On the other hand, some
corrosion may occur above the water table in disturbed soil and it is then
necessary to apply a protective coating. Normally two coatings of a bituminous
paint are given to steel sheet piles before installation.
Sheet piling about an excavation frequently requires support. In such cases
sheet piling can be driven in advance of excavation. As excavation proceeds,
waling is placed against the sheet piling and struts are placed across the
excavation and wedged against the waling. Alternatively, soldier beams are
driven at intervals along the line of excavation, and as the excavation proceeds
wooden sheeting planks are inserted horizontally against the ground and are
supported by the soldier beams. Tiebacks or anchors also may be used as
support.
Steel sheet piling is used most frequently for anchored retaining walls. These
normally range in height between 4.5 and 12 m, depending on the soil conditions.
Because of its structural strength, watertightness and ability to be driven to
appreciable depth in most soils, steel sheet piling is also widely used for
cofferdams. However, in some soils—notably fine sands and silts beneath the
water table—there is a danger of quick conditions developing if the critical
gradient is exceeded as ground is removed. Piping sometimes occurs along steel
sheet piling, especially when it is driven with the aid of jetting. In such situations
consideration should be given to lowering the water table below the base of the
excavation.
2.2
CONTIGUOUS BORED PILES AND SECANT PILES
Contiguous bored piles are frequently associated with both shallow and deep
excavations for basements to buildings and with cofferdam work. This is
particularly the case in urban areas where noise or the effect of installation on
adjacent property are important, as well as in industrial complexes where access,
headroom or restrictions on vibration may make other exclusion methods such as
steel sheet piling less acceptable.
Contiguous piles can be designed to cantilever vertically, to be strutted or to

be anchored back. It is usual to employ ground anchors, either temporarily or as
part of the final design, to eliminate the need for struts which limit the working
space within an excavation and also to reduce the required pile diameter and
amount of reinforcement within the pile. Bored piles possess a number of
advantages. For example, their length can be varied to suit the ground
conditions. Normally temporary casings have to be used during construction of
contiguous bored piles although in clay soils only a short length of casing
generally is necessary to provide a seal at the top of the clay. Contiguous bored
piles normally are best suited to cohesive soils. By contrast they may be
impossible to install in saturated loosely packed silts or sands owing to the soil
undergoing liquefaction. Moreover a loss of ground may occur when sinking
piles in groups in non-cohesive soils, with possible settlement of adjacent
structures. In soils where the boring is slow or difficult, owing, for example, to
the presence of boulders, the amount of overbreak resulting from driving and
boring out casings at very close centres in line, is likely to be greater than that
arising from single piles at wide spacing. Difficulties can arise when constructing
bored piles in water-bearing or in squeezing ground. For instance, if the bottom
of the casing is lifted above the base of the concrete pile while concreting
operations are still proceeding, with the casing being withdrawn in a number of
lifts, then water may enter the borehole and thereby lower the strength of the
concrete. Under artesian conditions water may pipe up the pile shaft and in so
doing remove cement. Squeezing ground may give rise to ‘necking’ in a pile.
When contiguous bored piles are used for retaining walls, the piles are
installed either in a single or double row and are positioned such that they are
touching or are in very close proximity to each other. The piles are reinforced to
resist the applied bending moments and shear forces. There is much flexibility as
far as the choice of pile diameter is concerned. However, the circular section of
the pile generally means that more main reinforcement per metre run of wall is
required than in the constant section of a diaphragm wall in order to produce an
equivalent moment of resistance (North-Lewis and Lyons, 1975).
Presetting a number of casings in advance of the piling ensures the correct
spacing of the piles in the line. The depth to which the casing extends depends on
the ground conditions. For instance, where the piles are in clay, only a short
length of casing generally is necessary to provide a seal in the top of the clay.
Boring can then continue in an uncased hole. The use of temporary casing leads
to gaps between piles. Such gaps normally should not exceed 75 mm at the head
of the pile, and the verticality of the pile shafts should remain within the
commonly accepted tolerances for bored piles (for example, 1 in 75).
The concrete of contiguous bored piles is placed directly into the pile shaft
through a short trunking and hopper for dry conditions, and by the tremie
technique under water. Alternate piles are constructed and the concrete allowed
to harden before boring the intermediate piles. The holes are excavated by a
grabbing rig or power auger.
Where water cannot be removed from a borehole either by baling or pumping,

then the hole has to be cased from top to bottom to prevent collapse of the sides
and a concrete plug is formed, beneath the water, at the base of the hole. Water is
pumped from within the casing when the concrete plug has hardened
sufficiently. Prior to filling the casing with concrete, it should be gently turned
and lifted slightly to release it from the plug (Tomlinson, 1986).
When the concrete has hardened, the casing is then removed. Alternatively,
concreting under water from a tremie pipe can take place over the complete
length of the borehole. Such construction may not be able to avoid the formation
of weak zones in the pile.
Where contiguous bored piles pass through water-bearing strata various
grouting techniques can be employed to seal the joints. Grouting of joints is
carried out before the excavation begins. The watertightness of the structure can
be further improved by grouting the joints on the exposed face after excavation.
Watertightness also can be achieved by using interlocking secant piles.
Alternate piles are first installed, but at a closer spacing than in the contiguous
method (Figure 2.3(a)). Then the intermediate piles are installed by boring out
the soil between alternate pairs, followed by chiselling a groove down the sides
of the pile shafts. Lastly, concrete is placed in the holes, including the grooves,
thereby forming a fully interlocking watertight wall.
A more recent method of installing secant pile walls is to form alternative
bentonite-cement piles and reinforced concrete piles which interlock to provide a
continuous wall. The bentonite-cement piles are constructed first, then the
reinforced concrete piles are constructed midway between them. Parts of the
bentonite-cement piles are bored out in order to form the reinforced concrete
piles and in this way the interlock is achieved. The bentonite-cement piles are
constructed in two sequences. First, piles are formed at every second position
(i.e. at double their spacings). Then after the first sites have set a second row of piles
is constructed midway between the first. The reinforced concrete piles are
emplaced when the bentonite-cement sites have gained sufficient strength for
boring to proceed (Figure 2.3(b)). This is usually at least 36 hours after the
bentonite-cement piles have been constructed.
2.3
SLURRY TRENCHES
Slurry trenches are used extensively as a means of ground water cut-off. For
example, they have been used to control seepage beneath dams and to contain
groundwater pollution, notably from sanitary landfills and hazardous waste
impoundments. Slurry trenches can achieve permeabilities of the order of 10−9 m/
s.
There are two types of slurry trenches, namely, soil-bentonite and cement-
bentonite. In the soil-bentonite cut-off process, the bentonite slurry is displaced
by a soil-bentonite mixture similar in consistency to high-slump concrete. The soil-
Figure 2.3 (a) Sequence of installation of secant piles: (1) initial piles emplaced; (2)
groove chiselled in initial piles; (3) alternate piles cast. (b) Four stages in the construction
of a secant pile wall using bentonite-cement and reinforced concrete piles (after Anon.,
1985 (a)).
bentonite backfill consists of excavated soil, with the addition of material from a
borrow pit mixed with bentonite mud, and forms a low permeability, highly
plastic cut-off wall. In the cement-bentonite cut-off process cement is added to a
fully hydrated bentonite slurry. The addition of the cement causes the slurry to
harden, giving a strength comparable to that of stiff clay. At suitable sites, where
slurry trenches have been used for basement, precast concrete panels have been
lowered into slurry trenches and exposed by subsequent excavation to form the
permanent basement walls (Figure 2.4). Self-setting slurry is used.
Impermeability is the most important property of a cut-off wall. There are
three factors which account for the impermeability of a slurry trench, namely, the
‘grouted’ zone, the filter cake and the backfill. Bentonite initially may permeate
the pores of the adjacent soil, depending upon the size distribution of the pores.
As bentonite is thixotropic, when left undisturbed it gels and thereby seals the
pores. The distance to which the soil is affected can range from almost zero in
dense clays to a metre or so in loosely packed sands and gravels. A filter cake of
loosely packed bentonite particles forms after a relatively short period on the
sides of the trench during excavation; this results from the filtrate as the fluid
escapes into the surrounding soil. This filter is usually a few millimetres in
thickness and has a permeability as low as 10−11 m/s. Hence, it acts as a
watertight membrane. The nature of the backfill has been referred to above.
Permeability of the backfill can be as low as 10−9 m/s.
The chief factors affecting the permeability of a cement-bentonite slurry are the
cement/water ratio, the bentonite/water ratio and the mechanical procedure in
making panel connections between the fresh cement-bentonite slurry and the set
cement-bentonite. It is important that the bentonite be fully hydrated before any
cement is added. This means that the cement/water ratio is the factor which
controls the ultimate permeability of the backfill (generally between 1:4 and 1:
5). The permeability of a cement-bentonite slurry cut-off wall can be as low as
10−9 m/s. Activated sodium bentonites have large swelling potential and are
stable with cement. Between 35 and 45 parts bentonite mixed with between 160
and 220 parts cement, and water, according to Schweitzer (1989) gives a
‘plastic’ wall with an ultimate unconfined compressive strength of 1 MN/m2 or
over. This should allow a slurry wall to deform under loading without cracks
developing.
Because it must have a low permeability, the continuity of slurry trench
excavation is important. The depth of a slurry trench is controlled in many cases
by the ground conditions, notably the depth of an impermeable formation, and
the type of barrier to be constructed. The minimum depth of penetration into an
impermeable formation at the base of the cut-off wall depends upon the nature of
the formation. For instance, if this is competent impervious rock, then only a
very small penetration may prove satisfactory. On the other hand, an excavation
into clay may need to go to a depth of 1 m or so. The width is governed by the
required permeability of the cut-off wall, the head of water across the wall, the
size of the excavation equipment available, and the materials that form the wall.
Figure 2.4 Precast concrete panels placed in slurry trench to form part of basement wall
which is anchored back
However, because the filter cake provides the principal barrier to groundwater
movement, the width of the trench is not a major factor. In the case of soil-
bentonite backfill slurry walls, Millet and Perez (1981) suggested that the wall
should have thicknesses varying from 1.5 to 2.3 m. The increased shear strength
of the backfill forming cement-bentonite slurry trench cut-off walls allows them
to be thinner, their thickness ranging from 0.6 to 0.9 m. Such widths are
satisfactory for depths of up to 30 m of hydrostatic head. Beyond this the ability
of the cement-bentonite cut-off wall to withstand hydrofracturing should be
determined. The deviation from verticality is only important insofar as it affects
the continuity and integrity of the wall.
A major factor controlling the deformability of cut-off walls is the properties
of the materials of which they are composed. Soil-bentonite walls are quite
deformable and do not have problems with regard to cracking. The cement/water
ratio has a dramatic effect on deformability of cement-bentonite backfill. For
instance, the higher the cement/water ratio, the higher the strength, the more rigid
and hence less deformable is the eventual cement-bentonite wall. On the other
hand, the higher the bentonite/water ratio, the more flexible, and the more
deformable the wall may be.
Excavation is made under slurry using backhoes, clamshells, draglines and
special devices. The head of slurry should be kept to 1–1.5 m above the
maximum anticipated groundwater level in the trench. The method of excavation
chosen is based primarily on the width and depth of the trench, and the type of
soil. More than one method of excavation can be used, for instance, a dragline
can be employed to dig the upper part of a trench (down to around 21 m), greater
depths being removed by clamshell. Speed of construction is increased if several
alternate panels can be formed during the initial hardening period, the grab
returning to dig out the intermediate panels which, being of lesser width, are cut
into the primary panels. There are no joints between panels, therefore stop-end
tubes and guide trenches are not required.
As excavation proceeds, sands and silts become suspended in the slurry. When
a sample of slurry obtained from the trench exceeds a specific gravity of about 1.
6, the slurry should be circulated through settling ponds or separators. This
improves the quality of the filter cake on the sides of the trench and enables the
backfill to settle into position satisfactorily.
Specifications for soil-bentonite backfill have involved a wide range of
particle sizes from coarse to fine, with the bentonite content in the range of 2–4%
by weight, and of the order of 10–20% fines. Silty sand provides the ideal
material for this purpose. Clays are suitable, except that hard clays tend to remain
as blocks and therefore may give rise to voids in the fill. Cobbles and boulders,
as well as roots or other organic material, should be removed before placement
of fill.
The backfill is mixed with slurry at the surface, using bulldozers or front-end
loaders, before being placed in the trench. The initial backfilling operation for a
soil-bentonite cut-off wall involves the backfill being placed at the bottom of the
trench by a clamshell bucket. This continues until the backfill material reaches
the top of the trench. Then the remainder of the backfill at the surface is moved
by a bulldozer onto the backfill exposed in the trench. This action forces the
material to slide down into the trench under its own weight, and continues until
the entire excavation has been backfilled. There is always some consolidation
and settlement of a soil-bentonite cut-off wall. The effects of consolidation on
the backfill material at the top of the soil-bentonite cut-off wall may lead to the
development of seepage paths.
When slurry trenches are used to control groundwater pollution from
hazardous waste sites, it may be that some of the chemicals present are
incompatible with bentonite. In such situations high-density polyethylene
sheeting is placed in the trench prior to backfilling. Sheeting also can be used to
form an impermeable membrane when excavated material is not readily suitable
for use as backfill yet cannot be economically disposed of. The permeability of
high-density polyethylene sheeting can be as low as 10−14 m/s. When a slurry

trench cut-off is required with a permeability of less than 10−9 m/s, then such
sheeting is used to enclose the backfill.
Ressi and Cavalli (1984) described the construction of a geomembrane
sheeting-enclosed cut-off wall (the Environwall). First, a narrow trench is
excavated and filled with bentonite slurry to prevent the trench collapsing. The
geomembrane sheeting is fabricated to form a U-shaped envelope fitting the
dimensions of the trench and then ballast is placed within it so that it can sink
into the slurry trench. After initial submergence into the trench, the envelope is
filled with wet sand. A system of wells and piezometers can be installed in the
sand to monitor water quality and piezometric pressure, respectively. If the
sheeting is damaged, then the system will detect the leakage. Wellpoints can be
used to abstract this water and to maintain the net gradient into the liner.
In the beam method a cut-off wall is constructed by forming a narrow cavity in
the soil by repeatedly driving or vibrating an I-beam section within it. As the I-
beam is extracted an impervious material, usually a mixture of bentonite and
cement, is pumped into the cavity (Figure 2.5). Successive penetrations of the I-
beam are overlapped to develop a continuous membrane (Boyes, 1975). In the
driven pile system, a clutch of H-piles is forced into the soil along the line
of advance, the end pile being withdrawn and redriven at the head end. Grouting,
by a tube fixed to the web, is carried out during withdrawal. Walls are limited to
a maximum of 20 m depth with these systems due to the danger of deviating
during driving.
2.4
DIAPHRAGM WALLS
Diaphragm walls represent a method of constructing walls in the ground in a
narrow trench filled with bentonite slurry. Diaphragm walls compare favourably
in terms of watertightness, stiffness and mechanical strength, with cut-offs
formed of steel sheet piling, precast piles or cast in-situ piles. They may be rigid
(concrete) or plastic (concrete and bentonite mixture) when load-carrying
capacity is not required. Alternatively, diaphragm walls may be used as load-
bearing and retaining walls and in such cases they are reinforced by
incorporating a steel cage. Diaphragm walls are usually more economical when
used as part of the permanent load-bearing structure.
A diaphragm wall allows a small area to be isolated for excavation using the
wall as a sort of cofferdam. In such instances the panels are constructed to form a
polygon. Diaphragm walls are composed of panels that are formed either
successively or alternatively. In the first case, panels are formed next to each
other in line. In the second case, primary panels are formed at regular intervals
and the wall is then completed by constructing secondary panels between those
already in place. Obviously watertight joints are required. In ground where
artesian pressure exists, even a slight tilt of the panels relative to each other or
Figure 2.5 Beam method of forming a cut-off by driving or vibrating an I-beam. The length
of the I-beam has to be at least equal to the depth of the cut-off being constructed
inadequate concreting could cause disastrous flooding in deep excavations. This

is even more important if the situation is aggravated by tidal fluctuations.
Diaphragm walls possess several features suitable for use in built-up areas
(Boyes, 1975). For example, they can be constructed immediately adjacent to
columns subject to live loadings, to multi-storey buildings and to old
foundations. Structural waterproof walls may be formed in advance of the main
excavation, which permits above- and below-ground construction at the same
time if necessary. In many applications the interior portions of concrete
diaphragm walls can be left exposed, as in subway excavations or basement car
parks. Furthermore, the technique is virtually vibrationless and quiet, and so

offers minimum disturbance.
The subsurface conditions should be thoroughly investigated so that any
potential problems during excavation attributable to the presence of pervious
layers are revealed. In particular, the ground water conditions around the trench
should be understood. This includes seasonal variations, as well as potential for
dramatic changes in groundwater levels caused by anomalous weather
conditions. Artesian pressures, springs and artificial sources of water (such as
potential broken sewers) should be recorded to avoid instability of the trench
during excavation.
In the construction of a diaphragm wall, a trench is first excavated to a depth of
a metre or so. The guide walls are then constructed. The guide walls enable
excavation of the main trench to begin below levels where made ground, drains
or footings and other surface obstructions might be met. They support the trench
against heavy construction surcharge pressure at levels over which the slurry rises
and falls, and protect the sides against wave action set up during digging or
during the introduction of fresh slurry.
A panel of predetermined dimensions is excavated between the guide walls
(Figure 2.6). The width of a panel usually varies between 0.5 and 1 m, although
greater widths are excavated when required. Lengths of up to 5 or 6 m represent
the maximum that can be concreted conveniently with one tremie pipe (Millet
and Perez, 1981). Problems of supply and coordination sometimes arise when
more than one tremie pipe is used.
Slurry is introduced into the panel and maintained at as high a level as possible
without spilling onto adjacent ground, thereby offering maximum support to the
sides (Anon., 1985(b)). Individual panels are excavated by first digging to the
full depth at each end, then the central section is excavated. Each grab or drill
excavates a full panel width. The depths required in most structural work present
little difficulty. The unavoidable fluctuation of the level of the bentonite in the
trench during digging must be taken into account. It is generally recommended
that the level of the slurry in the trench must be at least 1.25 m above that of the
surrounding water table. If this cannot be realized, then the trench may be dug
within a raised embankment, the water table may be lowered by installing
wellpoints or the slurry may be weighted by adding sand or powdered barite.
After the panel has been excavated, any slurry-soil sediment that has
accumulated at the base of the trench is removed and replaced with fresh slurry or
treated with chemicals. Frequently, cleaning the base of the trench by removing
the sludge with a grab proves adequate.
In a reinforced diaphragm wall the reinforcement cage is lowered into place
through the slurry and supported from the concrete guide walls of the trench.
Reinforcement is made up into cages prior to placement.
The concrete is placed using a tremie pipe. Slurry is pumped from the trench
as the concrete is placed, being removed at a similar rate. Pours should be
completed in the minimum time (up to 30 m3/h) and with minimum discontinuity
in order to avoid embedment of stop-end tubes, blockages in the tremie pipe and
flotations of the cage due to upward drag from stiff concrete. All the concrete for
a panel must be poured before setting or significant stiffening occurs. In practice
this generally means 3–3.5 h.
As the concrete for each panel is tremied in, it moulds against the concave end
of the previously cast panel at one end and against a circular steel stop-end
placed temporarily at the other end. The knuckle-joint between adjacent panels
provides sufficient connection for normal diaphragm wall work. Nevertheless,
this simple butt joint between panels cannot prevent water penetration; however,
significant leakages are rare. This is probably due to the presence of a thin layer
of contaminated bentonite at the edges of the joint. Where leaks take place they
are associated with differential deflections between wall panels, and these are
worst near corners.
Because bentonite cannot penetrate impervious clays, filter cake does not form
in such ground conditions, hence the suspension is totally displaced by concrete
and a perfect soil-concrete contact is obtained. Conversely in pervious sand or
gravel, all the free bentonite is displaced but the advancing concrete cannot
displace the soil impregnated with bentonite. Furthermore, the harder bentonite
filter cake is not removed, nor any soft filter cake that occurs in cavities.
For practical purposes concrete used for diaphragm walls can be regarded as
impermeable. Because of the humid conditions in which curing takes place,
shrinkage cracks due to drying are almost eliminated.
The commonly accepted tolerance on verticality is 1:80. However, the control
that can be achieved is governed by the soil, the plant and the early detection of
any deviation. Obstructions such as boulders or hard strata, especially if inclined,
increase the difficulties appreciably and correction normally involves the use of
long, heavy chisels. Millet and Perez (1981) recorded that walls have been built
to depths exceeding 122 m with less than 150 mm deviation from the vertical, a
tolerance of 1:800. However, the depth limit of diaphragm walls excavated with
grab equipment usually ranges between 40 and 50 m since, beyond these depths,
such equipment tends to deviate beyond the vertical tolerance and therefore can
offer poor guarantee on the quality of the joints and, hence, the continuity of the
wall. The use of percussion tools with direct or reverse mud circulation allow
this depth limit to be exceeded, but their production rate is low (De Paoli et al.,
1989). The slurry trench cutter (see below) is a new type of equipment that can
excavate diaphragm walls to more than 100 m and guarantee their continuity.
Goto and Iguro (1989) also described the construction of a deep wall with
deviation well within tolerance. The drilling unit of the excavator had an
adjustable guide and excavation was done by wing bits (Figure 2.7).
During the excavation of a bentonite trench sections of the soil walls, even if
only for a short period, have to be free-standing prior to the development of the
filter cake. Free-standing height does not present a problem in cohesive soils but,
by contrast, in dry coarse-grained soils lacking interlocking particles and surface
Figure 2.6 Construction of a reinforced diaphragm wall. General procedure for

excavation of panels during which time the excavation is kept filled with bentonite
suspension; (a) first one end, then opposite end of panel is excavated to full depth; (b)
third and last stage is the excavation of the centre panel. General procedure for concreting
panels when the bentonite is displaced by concrete: (c) steel stop-end pipes and
reinforcement cage positioned and concrete placed through tremie pipe; (d) section
through completed panel showing guide walls and steel; (e) illustrating variety of uses of
steel stop-end pipes as required
tension forces due to moisture content, and with no osmotic force between slurry
and soil, it is impossible to achieve free-standing height (Veder, 1984).
Figure 2.7 Guide wall and excavator (after Goto and Iguro, 1989)
If excavation work is not carried out properly, then soil movement can occur
about the trench. Sliding or caving of the sides with associated slumping of the
guide walls and subsidence at the surface may result. Such problems are more
likely to occur in coarse- than fine-grained soils. In dry coarse-grained soils they
can be avoided by grouting or compacting before excavation begins. If coarse-
grained soils are moist, then sufficient time should be allowed for the
development of the filter cake, and the rate of excavation should be slower than
usual.
It may prove impossible to construct diaphragm walls by normal methods in
soils with very low strengths. This is because the internal pressure of the
bentonite occupying the trench may be less than the active pressure of the
adjacent soil. For example, instability commonly occurs in soft marine clays. In
fact, as Sliwinski and Fleming (1975) pointed out, stability is also influenced by
the length and shape of the panel and arching effects. They recommended that
any soil in which =0 and the cohesion is less than 10 kN/m2 should be treated
cautiously. In such cases panels should be short and uncomplicated.
Another disadvantage concerning the construction of a diaphragm wall is a
high water table. This hinders the casting of guide walls. It also reduces the
differential between the pressure inside the trench and the active pressure of the
adjacent soil. Not only does this have a detrimental effect on the stability
conditions but it also adversely influences the formation of filter cake.
If excessive loss of bentonite occurs in pervious strata, especially if the

viscosity and gel strength are low, then the formation of filter cake may be poor
or non-existent, or may take place only slowly. Viscosity and gel strength may
not be adequately developed if the bentonite suspension is not fully hydrated. In
fact the rate at which the trench is excavated should take account of the caking
properties of the slurry as well as the strength of the soil. A number of measures
can be taken to counteract the loss of slurry into adjacent soil, including
thickening the slurry by adding further bentonite or cement. Alternatively, the
trench can be backfilled with lean concrete, which is excavated shortly after it
has started to set. Such measures must be taken immediately otherwise the sides
of the trench will begin to collapse as the level of the slurry sinks.
A diaphragm wall was constructed around the site at Sizewell B nuclear power
station in Suffolk. It was sunk to a depth of approximately 55 m into the London
Clay. Initially, dewatering was considered to lower the water table within the site
but reservations were expressed regarding possible settlement that might have
affected Sizewell A station. Other considerations taken into account were the
prevention of local wells and watercourses drying up and the avoidance of saline
pollution of the neighbouring marshes. The plastic concrete cut-off wall has a
minimum compressive strength of 1.5 N/mm2, an in-situ permeability of 1×10−12
m/s and is 800 mm thick. It contained 50% PFA and the wall can flex with the
surrounding ground. Along the foreshore the wall was cast in reinforced concrete.
A new digging unit (Figure 2.8), developed for deep excavation in granular
soils or very compact ground, was used (Anon., 1988). The digging machine,
suspended from a crawler crane, cuts a vertical slot to the required width and
disposes of the excavated material via a hydraulic reverse mud circulation system.
Excavation started by trenching to a depth of 3 m with a crane-mounted grab,
under bentonite. This provides sufficient depth to submerge the cutters and pump
of the digging unit. Before starting excavation with the digging unit a steel guide
frame is installed between the guide walls. Each primary panel was 6 m long.
Secondary panels in the plastic concrete sections also were 6 m in length but
those for the reinforced concrete wall were only 2.4 m. Desanding took place
after each panel was excavated. The panel was then concreted by the tremie
method. The secondary panels were cut into the concrete of the adjacent primary
panels, thus eliminating the need for tube-joints or end-stops to form a watertight
seal. It could take 7 days for the plastic concrete to gain enough strength to allow
excavation of secondary panels. The required strength was gained much more
quickly in the reinforced concrete panels.
Hansmire et al. (1989) monitored lateral movement and settlement of a
concrete diaphragm wall and found that the movements were small, only a few
millimetres. Tie-backs were used as a temporary support for the wall until the
associated structure was completed. Their monitoring programme allowed
Hansmire et al. to make several recommendations for design. One was that if tie-
backs are used, their long-term benefit to the final structure should be considered
in relation to reducing reinforcement of long spans. Another recommendation
Figure 2.8 Diagrammatic outline of the Soletanche Hydrofraise rig
was that, in relatively stiff soil in which the strain is kept to well below that
where creep movements could develop, the lateral loads assumed in design
probably will not be realized. On the other hand, where soils are very soft, long-
term lateral loads should be considered.
2.5
THE CHARACTER OF BENTONITE SLURRY
Bentonite slurries, which are used to give support to excavations, can vary
widely in their properties; however, according to Hutchinson et al. (1975) they
must:
1. support the excavation by exerting hydrostatic pressure on its walls;

2. remain in the excavation and not flow into the soil;
3. suspend detritus to avoid sludgy layers building up at the base of the
excavation;
4. be easily pumped;
5. be capable of clean displacement by concrete, with no subsequent
interference with the bond between reinforcement and set concrete;
6. be able to undergo screening or hydrocycloning to remove detritus and
enable recycling.
Usually, the first three items require thick dense slurries, while the latter three
are best achieved using very fluid slurries. As a consequence, there is a conflict
of requirements which has to be resolved in order that an acceptable specification
regarding slurry properties can be developed.
One of the basic functions of a bentonite slurry is to form a filter cake layer on
the exposed soil as the trench is excavated. The filter cake only develops if the
fluid pressure in the trench exceeds any external groundwater pressure and if the
permeability of the soil is within certain limits (Sliwinski and Fleming, 1975).
The bentonite slurry is prepared by adding a measured quantity of the powder
to clean fresh water. It is well mixed and allowed to hydrate before it is
introduced to the trench. Mixing and hydration are necessary if the slurry is to
form a good filter cake. The time taken for hydration, after which the bentonite
suspension has attained the desired rheological properties, depends on the
method of mixing. For instance, a strong shearing action leads to rapid hydration
and good performance.
If a bentonite slurry is to exert a stability pressure on a trench with permeable
walls, then it must seal the surface with which it comes into contact. This
prevents slurry penetrating into the soil.
Surface filtration occurs when a filter cake starts to build up by hydrated
particles of bentonite bridging across the entrances to pores in the soil, without
penetrating the soil to any great extent. Water continues to percolate from the
slurry through the filter cake into the soil during and after the filter cake has
formed.
Deep filtration occurs when a slurry slowly penetrates soil, blocking the pores
and developing a filter cake within them. In such instances the seal may extend
several centimetres into the soil. The concentration of bentonite in the filter cake,
in surface as well as deep filtration, is greater than in the slurry (typically 15% for
a slurry containing 5% bentonite).
In this context, determination of the swelling period—that is, the time that
elapses between the end of mixing and the introduction of the slurry into the
trench—is important (Veder, 1984). For example, when used in highly
permeable soils the swelling period should be short to enable the suspension to
continue to swell in the trench, thereby preventing deep penetration into the soil
with accompanying significant loss of bentonite.
The concentration of bentonite in a slurry has to take into account the often
conflicting requirements of trench support, prevention of losses to adjacent strata,
reduction of drag and flotation of the grab during digging and the need to
prevent too much soil being suspended in the slurry. Normally the concentration
of bentonite in slurry varies from about 3% in stiff clays to about 8–10% in dry
open ground. For instance, a 4–6% concentration is retained by soil with a
permeability up to about 10−5 or 10−6 m/s. If the permeability of the soil exceeds
this, then a denser suspension, ranging up to about 12%, may prove necessary. In
exceptional cases, where even this dense fluid is not retained, a number of
additives may be tried. However, none of these methods is likely to be very
effective if permeability exceeds about 0.5−3 m/s. In open ground (k=0.01 m/s),
the sealing mechanism can be changed by including small quantities of fine sand
in a slurry, with a resultant significant lowering in the initial loss of fluid. Bentonite
entering the trench should contain less than 1% fine sand to give a good
performance. As far as the formation of a filter cake is concerned, the presence
of fine sand can counteract, to a certain extent, incomplete hydration. The loss of
slurry to open dry ground can also be avoided by increasing the viscosity by
adding cement (up to about 17 kg per cubic metre of slurry) or by the addition of
sawdust, cellulose flakes or shredded clay.
The presence of water in the ground adjacent to a trench can impede the
penetration of the surrounding strata. Consequently, slurry concentrations may
have to be reduced to a level at which they provide sufficient support to the sides
only, usually to between 6 and 8% bentonite concentration. Salty and limey
water cause slurry to thicken and, in extreme cases, may lead to bentonite settling
out of suspension. In practice, however, these have not proved serious problems.
They can be taken care of by adding a phosphate-water mix (up to 1%
phosphate) to thin the slurry to its original viscosity. Ground consisting of chalk
can give trouble by reducing the effective concentration of bentonite during
excavation, thereby diluting the slurry. Peaty water also has the effect of thinning
the slurry. In such cases it may be necessary to increase the bentonite
concentration to restore the viscosity. In fact, the presence of organic compounds
can render the method impracticable.
The pH value is particularly significant where the chemistry of the soil

excavated or of the groundwater can give rise to a dramatic change in the pH
value of the bentonite slurry. Millet and Perez (1981) proposed that the most
desirable range of pH value of slurry is of the order of 6.5–10. If the pH value
exceeds 10.5, the clay particles in the slurry will have a tendency to flocculate
and settle out. If this happens, then a deflocculating agent has to be added to
ensure continued effectiveness of the slurry.
From the point of view of wall stabilization, the higher the density of the
slurry the better. The density of freshly prepared slurry varies from about 1.02 t/
m3 for a mix with 4% bentonite to 1.05 t/m3 for a mix with an 8% bentonite
concentration. According to Hutchinson et al. (1975), the minimum allowable
density should be 1.034 t/m3. In other words, it is slightly above that of the
groundwater. Millet and Perez (1981) maintained that specified maximum
density typically is of the order 1.04 t/m3. This is to ensure that the tremie
process is not impeded. In fact, if slurry is to be displaced by tremied concrete its
density should not exceed 1.3 t/m3 and its plastic viscosity should not be above
20 cP* (i.e. 2×10−2 N s/m2; equivalent to a maximum bentonite concentration of
about 15%). Fine sand is added to increase the density of a slurry. However, the
amount should be limited, again so that the subsequent tremie concreting
operations are not impeded. For example, it is generally specified that the sand
content should not exceed 5%.
During excavation detritus inevitably becomes mixed with the slurry. Veder
(1984) remarked that the incorporation of about 5% fine particles in the slurry
does not impair its effectiveness but a higher percentage leads to a number of
shortcomings. Not only does incorporation increase the average density of the
slurry, which makes excavation by the grab in the slurry more difficult, but it can
generate slowly settling layers of sand and silt. These, in turn, produce density
gradients in the slurry and lead to a build-up of sludge in the bottom of the
trench. Unfortunately this means that tremied concrete is unable to displace this
material from the base of the excavation. Moreover, such slurry tends to have a
high viscosity, with adverse effects on displacement of bentonite by concrete
past a reinforcing cage. Indeed incomplete displacement of bentonite by concrete
can give rise to the formation of cavities.
If the amount of detritus in the slurry slows down the rate of excavation
significantly, then the slurry must be replaced. It certainly must be replaced
before the reinforcement is placed or concrete poured. If the bentonite slurry is to
be re-used, then it must have sandy or silty material removed.
By far the most common form of contamination is by cement, which occurs to
a varying degree during tremieing. Contamination by cement leads to the
formation of thick permeable filter cakes, very high fluid losses and high
viscosities. Hutchinson et al. (1975) showed that, in a standard bentonite slurry,
* IcP=0·001 N s/m2.
the effects become noticeable at 0.3% cement contamination, which is the

equivalent to a slurry with a pH value of about 11.7. They therefore proposed that
this figure should be the maximum pH value of a slurry.
2.6
COMPRESSED AIR
Compressed air can be used in tunnels, shafts and caissons to prevent the inflow
of water and to stabilize the ground exposed by excavation. In soft clay the
pressure of the air provides some support, but in sands the air drives the water
back from the exposed face so that capillary attraction between the moist grains
helps to stabilize the ground. Obviously the permeability of the ground
influences air loss and, therefore, whether or not the method can be used.
However, compressed air is a health hazard and its use should be avoided where
possible. Another drawback is that in large subaqueous tunnels such as vehicular
tunnels it is difficult to select an air pressure that will minimize water flow at the
invert, where the water pressure is naturally higher, without causing blow-outs at
the soffit.
3
Ground freezing
3.1
INTRODUCTION
Ground freezing involves the artificial lowering of ground temperatures so that
pore water is converted into ice, thereby reducing the permeability and
increasing the strength of the ground so treated. In this way an ice wall can be
formed around an excavation until a permanent structure has been constructed.
Frozen walls are frequently used in shaft sinking (Auld, 1985, Klein, 1989) and
to a lesser extent in tunnel construction (Harris, 1989; Floess et al., 1989). The
design of a frozen wall is basically a problem in heat transfer and, in the absence
of moving ground water, transfer in the ground is by conduction. The thermal
conductivity (Figure 3.1) and the heat capacity of the ground governs the rate at
which freezing proceeds. Fortunately the values of thermal conductivity for all
types of frozen soil fall within quite narrow limits. This is one of the reasons why
artificial ground freezing is a versatile technique, being able to deal effectively
with a great variety of soil types (Jones, 1981).
Artificial freezing can be employed as an exclusion technique to stop the flow
of groundwater into excavations, thus producing dry working conditions. In this
context, frozen walls can be used in conjunction with bored pile walls or slurry
trenches. The frozen ground provides a clear area that is free of bracing and earth
supports within which excavation can occur.
Ground freezing is influenced by the geological and hydrogeological
conditions at the site, the soil types present, and their thermal properties and
water content. Thermal conductivity varies vertically with each change in water
content and lithology. It also changes horizontally from the frozen to the
unfrozen ground. Obviously the position of the water table and its fluctuation are
important, as are the temperature and velocity of groundwater flow. Although
complete saturation of the soil is desirable, freezing may be carried out safely
when the degree of saturation is as low as 10%. Freezing of pore water in
cohesive soils never reaches completion; and attempts at ground freezing may be
doomed to failure if there is significant groundwater flow.
GROUND FREEZING 53
Figure 3.1 Thermal conductivity of soil (from Mitchell, 1981)

The mechanical properties of frozen ground are much more dependent on the
time taken in freezing and the ground temperature than on the nature of the soils
involved. Consequently, artificial ground freezing is less sensitive to geological
conditions than other methods of ground treatment and may be used in any moist
soil, irrespective of its structure, grain size or porosity (Jessberger, 1985). Weak,
running or bouldery ground, and ground in which the pore-water pressures are
high, lend themselves to freezing. Soil profiles, too heterogeneous to grout
predictably, are also amenable to freezing. However, in heterogeneous soils the
frozen zone tends to be irregular in shape. For a certain refrigerant temperature
the frozen ground is usually thinner in silts, clays and organic soils than in sands
and gravels (Figure 3.2). As these commonly are the weaker strata, design may
be dictated by the former three soil types. When clay soils are frozen, vertical
cracks may form in the soil between the frozen columns and the soil may bulge
near the zone being frozen. The fibrous structure of organic soils tends to be
altered permanently by freezing.
Figure 3.2 Content of unfrozen water in frozen soil (after Frivik and Thorbergsen, 1981)
3.2
METHODS OF GROUND FREEZING
Artificial freezing of ground is usually carried out in two stages, referred to as
the active and passive stages of freezing. Active freezing involves freezing the
ground to form the ice wall, while passive freezing is that required to maintain the
established thickness of the ice wall against thawing. The refrigeration plant has
to operate at a much higher capacity during the active than the passive stage of
freezing.
The primary plant with pumped loop secondary coolant is the system most
frequently used to freeze ground (Figure 3.3). An ammonia or freon refrigeration
plant provides the primary source of refrigeration. The system is worked by
either electricity or diesel motor and the condensers may be air or water cooled.
These plants have a wide range of capacities. Two stages are required to produce
temperatures below −25°C. Several different types of coolant have been used
with this system; however, the most common type of refrigerant is calcium
chloride brine. Other brines have been made from chlorides of sodium,
magnesium or lithium. The crystallization point of the chosen brine should be at
GROUND FREEZING 55
Figure 3.3 Ground freezing with supercooled brine

least 5°C lower than the ultimate temperatures at which ground freezing will
proceed. Typical brine temperatures for commercial refrigeration plants are
around −20°C to −40°C. A temperature of −5°C can normally be attained
throughout the zone to be frozen in reasonable freezing times and freeze-pipe
spacings (Sanger, 1968).
Heat is removed from the ground by way of probes placed in boreholes. The
installation of the freeze probes represents the most costly, time-consuming and
risky part of a ground-freezing project. Each probe consists of an external pipe,
closed at the lower end, and containing an open-ended inner tube of slightly
shorter length (Figure 3.4). Cooling fluid is normally introduced through this
inner tube. The coolant emanating from the bottom of the inner pipe is warmed
by the abstraction of heat from the ground as it moves upwards. Large quantities
of coolant have to be circulated to bring about freezing. The supply and return
temperatures of the coolant in the freeze probes are monitored to determine the
amount of heat removed and whether areas are being under- or over-cooled. Such
situations can sometimes be corrected by varying the flow of brine. The
temperature gradient along a freeze probe is greater when the brine is circulated
slowly. The refrigeration capacity of each freeze probe can only be increased by
increasing the capacity of the entire system. However, the ability to increase the
capacity of individual probes is desirable in order to control localized conditions,
especially unexpected ground water flows.
Figure 3.4 Ground freezing layout for an excavation

The growth of ice is more rapid at the base of the freeze probes. This can be
overcome by circulating the coolant more quickly or by installing a ring line
about 0.3 m below the ground surface immediately inside the perimeter of the
freeze pipes, thereby forming a beam of ice.
A warmer brine is discharged from the head of each probe into a collection
main, whence it is piped back to the refrigeration plant. There it is pumped
through a chiller and is delivered, re-cooled, via a distribution main to the inner
tubes of the freeze probes. Such a system, with two compressors, can ultimately
lower the brine temperature to −40°C if required.
All piping above the surface in a freezing system must be insulated to
minimize heat loss. Insulation is also necessary in large shallow excavations. A
temporary cover is desirable for large excavations which may be open for
months. However, in a deep excavation of small diameter such as a shaft, the air
is cooled appreciably and, as a result, protection of the exposed surface of the
frozen soil may be unnecessary.
Expendable refrigerants are very attractive for individual projects which only
last a few days or for projects where the cost of delay is high (for instance, for
emergency situations where rapid formation of a frozen wall is required or where
particularly low temperatures are needed to deal with heavy groundwater flow).
Although no refrigeration plant is required, the cost of refrigerant is high. In
expendable refrigerant systems the refrigerant is lost to the atmosphere after it
GROUND FREEZING 57
Figure 3.5 Ground freezing with liquid nitrogen
has absorbed energy in the ground and vaporized. Liquid nitrogen (LN2)
provides the fastest and thermally most efficient means of freezing the ground
(Figure 3.5). It evaporates at a temperature of −195°C at atmospheric pressure
and in doing so draws 162.4 kJ/l from its surroundings. About the same amount
of heat is needed to raise the cold gas to ambient temperature. In fact such
systems can sometimes freeze the ground in a matter of hours. The use of liquid
nitrogen also gives rise to fewer ice lenses, to less heaving and to greater strength
than freezing with brine. The use of liquid carbon dioxide is thermally less
efficient and normally is more difficult to control.
A tanker, from which the liquid nitrogen is injected into the freeze pipes, may
be used on small or emergency projects, otherwise one or more storage tanks are
needed on site (Gallavresi, 1985). Liquid nitrogen is introduced directly into the
circuit, by utilizing the pressures developed in the storage tanks, through a
distribution pipe carrying the liquid nitrogen to the freeze probes. These are
grouped and connected in series. There is a valve downstream from the terminal
freezing pipe through which the nitrogen gas is vented to the atmosphere. The
Figure 3.6 Required freezing time versus distance between freezing pipes (from
Jessberger, 1985)
pressure and consumption of liquid nitrogen are regulated at the storage tank in
accordance with the desired rate of heat extraction to form the frozen wall.
It is possible to combine brine and liquid nitrogen freezing. In this way the
freezing capacity can be raised to meet peak demands, or liquid nitrogen can be
used as a standby in the event of failure. Another possibility involves freezing
the ground with liquid nitrogen and then keeping it frozen with cooled brine.
Liquid nitrogen can be very quickly re-applied if unforeseen difficulties arise.
Most ground-freezing contractors have evolved their own method of
calculating the thickness of a frozen wall. When the thickness of the frozen wall
has been decided, the number of freeze holes and their spacing together with the
refrigeration capacity is ascertained. The most important factors governing cost
are the size and spacing of freeze pipes (Figure 3.6). Estimation of the time
required to freeze a radius R by using freeze pipes with radius ro can be obtained
from Figure 3.7. Relative spacings, R′, in excess of 15 normally are not used.
Shuster (1972) pointed out that without the correct field control of the liquid
nitrogen freezing process, the potential time savings shown in Figure 3.7 may
not be obtained. The energy requirement needed to freeze soil in kcal/m2 is
approximately 2200 to 2800 times the percentage natural moisture content.
A frozen ground-support system is a massive relatively rigid structure, which
is formed generally before excavation commences. Because of the relatively high
compressive and low tensile strengths of frozen soils, curved arched walls,
GROUND FREEZING 59
Figure 3.7 Determination of required freezing time (after Shuster, 1972, with permission
of ASCE). Indicated bands represent normal range of observed field and laboratory
results. However, results with forced convection of LN2 may vary more widely than
indicated due to the variables in control of the freezing process
particularly circular walls, provide the best structural solutions and should be
chosen whenever possible. Although a gravity wall may be constructed in a
straight line, the volume of ground that must be frozen varies from two to more
than five times as much as that required for a comparable curved wall.
3.3
HYDROGEOLOGY AND THE NATURE OF FROZEN
SOILS
Hydrogeology is one of the most important factors in an evaluation of the
feasibility of ground freezing, notably because of the heat contained in moving
ground water. In addition, the position of the water table plays a significant role
in ground freezing because of the importance of water content as far as the
mechanical properties of the frozen ground are concerned. Some coarse-grained
soils that are unsaturated may remain pervious when they are frozen. Hence, in
such a situation if the water table rises outside of the excavation after freezing
has been completed, water may seep through and cause thermal erosion of the
frozen wall. Nonetheless, as mentioned above, ground freezing may take place
when the degree of saturation is as low as 10%. Even when below 10%, the
moisture content of, for example, sands can be increased by injection of water
from the surface (Gonze et al., 1985).
If groundwater is saline or contaminated this can lower the point at which
freezing takes place and so reduce the strength of the frozen ground. Where
dissolved salts are present in the groundwater, chemical analysis should be made
and the temperature at which freezing occurs ascertained. Sea water presents no
special problems in this respect since its freezing point is only about 3°C below
that of fresh water. However, when a volume of saline water in granular soils has
been confined by a perimeter array of freeze probes, inward advance of the
frozen wall causes a progressive increase in the salinity of the unfrozen fluid. In
narrow excavations the salinity of the water in the unfrozen core is commonly
several times that of the original groundwater. Artificial ground freezing has
been used in formations containing natural brines derived from evaporitic
deposits. In such cases the temperature of the coolant must be much lower than
when groundwater is fresh, for example, it may have to be lowered to −20°C or
more.
The ground beneath the base of excavation is extremely important. Ideally, the
base of a frozen wall should be located in an impervious layer which forms a
seal, preventing the incursion of groundwater into the excavation. Where a
frozen wall is not founded in an impervious horizon, an excavation has an open
base. Consequently, extreme caution must be exercised to minimize water
movement beneath the frozen wall and to avoid piping or heaving of the floor.
Movement of water normally leads to an uneven development of an ice wall in
the direction of the flow. In addition, if the velocity of groundwater flow is too
great, the frozen soil columns will not merge, thereby leaving windows in the wall.
Obviously it becomes impossible to freeze soil if the amount of heat conveyed to
the freeze zone by flowing groundwater exceeds that being removed by the
freezing system.
It is generally accepted that the normal freezing process is not likely to be
successful where the flow of water exceeds 1.5 m/day. In such situations special
provisions have to be taken either to lower the flow rate or to cope with the
excess heat. If groundwater flow is greater than 1.5 m/day but less than 3 m/day,
the spacing between freeze pipes should be reduced or a second row of pipes
should be placed on the upstream side (Braun et al., 1979). Lower refrigerating
temperatures can be used. Calcium chloride can be chilled to −40°C before
increasing viscosity has a significant effect on pumping resistance. If colder
GROUND FREEZING 61
temperatures are required, then alternative secondary refrigerants or circulation of

the primary refrigerant can be used (Harris, 1989). When the flow exceeds 3 m/
day, either the groundwater gradient or the permeability of the formation must be
reduced. This is brought about by grouting before or during the installation of the
freeze pipes, or by intercepting the flow with wells. Generally the latter method
provides a more reliable solution. Liquid nitrogen has been used successfully
where groundwater flows of more than 50 m/day have been encountered, but this
is expensive.
3.4
STRENGTH AND BEHAVIOUR OF FROZEN GROUND
The long-term strength and stress-strain characteristics of frozen ground depend
mainly on ice content, the orientation of ice crystals with respect to the direction
of applied stress, the temperature of the ice, and duration of loading (Figures 3.8
and 3.9). The frictional resistance of soil particles is also important and is
attributable to the same factors which influence the strength of unfrozen soils—
that is, particle shape, particle size distribution, density, porosity and soil
structure (Sayles, 1989). Friction usually is the dominant factor determining the
strength of coarse-grained ice-poor soils. The behaviour of ice-rich soils in which
few of the grains are in contact is dominated by the characteristics of the ice
present. However, the maximum resistance offered by the ice is not necessarily
mobilized at the same time or strain as that of the limiting frictional resistance of
the soil. The cohesion of frozen soils involves the molecular attraction between
particles and the degree of particle cementation, including cementation by ice.
Cohesion increases with decreasing temperature but tends to decline with time.
A major part of the resistance to stresses that are rapidly applied is provided
by the ice matrix. Under rapid loading the short-term strength may be 5 to 10 times
greater than that under sustained stresses, that is, frozen soils are susceptible to
large creep strength losses. The compressive strength in short-term tests at low
temperatures may be up to 20 MPa.
The more clay material present in a frozen soil, the greater is the quantity of
moisture that remains unfrozen. Even so, the unconfined compressive strength of
frozen clays undergoes a dramatic increase in strength with decreasing
temperature (Figure 3.10). In fact, Lovell (1957) indicated that it appeared to
increase exponentially with the relative proportion of moisture frozen. For
instance, when tested, silty clay showed that the amount of moisture frozen at
−18°C was only 1.25 times that frozen at −5°C, but the increase in compressive
strength was more than four-fold. Similar results were found by Kujala (1989).
By contrast, the water content of granular soils is almost wholly converted into
ice at 0°C. Hence, frozen sands and other soils with relatively large pore spaces
exhibit a reasonably high compressive strength only a few degrees below
freezing, and there is justification for using this parameter as a design index of their
performance in the field, providing a suitable factor of safety is incorporated.
Figure 3.8 Unconfirmed compressive strength and cohesion of frozen sand and silt in
relation to temperature and time (after Jessberger, 1985)
The influence of both time and temperature is very important as far as the
stress-strain behaviour of frozen soils is concerned. In particular, stresses and
strains in frozen soils induced by externally applied loads are not constant but
vary with time. Other factors which influence stress-strain behaviour include the
overburden pressure and pore pressure during freezing, the amount of frost
penetration, water supply and ice lensing. The time-dependent qualities of
strength and viscous behaviour play a decisive role in the design of frozen soil.
Stress concentrations, which occur at the grain contacts when a frozen soil is
loaded, bring about pressure melting of ice. Hence, differential water surface
tensions are produced which mean that unfrozen water migrates to regions of
lower stress, where it freezes. Melting of ice together with water migration
disrupt the texture of a frozen soil, leading to plastic deformation of the pore ice
and a rearrangement of soil particles. This is responsible for creep (Sayles, 1989).
With deformation, the soil becomes more densely packed, giving rise to an
increase in strength as a result of an increase in internal friction between grains.
There is a concurrent reduction of cohesion and the amount of unfrozen water
GROUND FREEZING 63
Figure 3.9 Stress-strain relationship of frozen sandy silt at T=−10°C: (a) t=10 min; (b)
t=24h (after Jessberger, 1985).
may increase. If the applied stress is less than the long-term strength of the
frozen soil, the weakening process is offset by that of strengthening. By contrast,
if the applied stress is greater than the long-term strength, the strengthening
process does not compensate for the weakening process and the rate of
deformation increases with time. This eventually leads to the failure of the frozen
soil mass. Long-term deformations are hundreds of times greater than the initial
ones, while the continuous strength is from 5 to 15 times less than the
instantaneous strength.
Figure 3.10 Compressive strength of frozen sand and clay

The structured unfrozen water in frozen soil is in equilibrium with the ice and
the amount of water present depends on the type of minerals comprising the soil,
their specific surface, the temperature, and the salinity of the water. The
resistance to stress and deformation of a frozen soil is reduced as the content of
unfrozen water increases.
The deformation behaviour of frozen soil is viscoplastic, and both stress and
temperature have significant influences on the strain at any time. The creep
curves in Figure 3.11 for a frozen silty clay illustrate these effects. The onset of
tertiary creep represents the beginning of failure. Creep is temperature and time
related for each type of soil. In practice, the evaluation of stability of frozen soil
masses, the prediction of creep deformation and the possibility of creep rupture are
complex problems. This is because of heterogeneous ground, irregular geometries,
and temperature and stress variations throughout the frozen soil mass. However,
where the period of unsupported exposure is short and the temperature below
freezing point large, the scale of creep relaxation generally is not significant.
Frost heave may occur in some soils when they are frozen, depending on their
permeability and in-situ confining pressures. Such movement is due to the
GROUND FREEZING 65
Figure 3.11 Typical creep curves for silty clay at various applied stresses and
temperatures (from Bell, 1975)
conversion of pore water to ice on the one hand and to secondary frost expansion
which results from migration of pore water to the freeze front on the other
(Kujala, 1989). However, if the soil drains at the same rate as the progress made
by the freezing front, expansion does not occur. Usually clean free-draining
sands and gravels are not susceptible to either type of frost expansion.
Significant movements are possible in silts and clays. Mettier (1985) described
the reduction of heave from a maximum of 105 mm to 5 mm on certain sections
of the Milchbruck tunnel by applying intermittent refrigeration during the
maintenance stage.
Further causes of movement associated with artificial ground freezing are a
consequence, firstly, of stress relaxation upon excavation and, secondly, of thaw
consolidation. If the soil was previously relatively dense, then the amount of
consolidation is small. This situation occurs in coarse-grained frozen soils
containing very little segregated ice. On the other hand, some degree of
segregation of ice is always present in fine-grained soils. For instance, under
such conditions the moisture content of frozen silts exceeds significantly the
moisture content present in their unfrozen state. As a result, when such ice-rich
soils thaw under drained conditions, they undergo large settlements under their
own weight. As ice melts and settlement occurs, water is squeezed from the
ground by overburden pressures or any applied loads. Excess pore water

pressures develop when the rate of ice melt exceeds the water discharge capacity
of the soil.
It has generally been assumed that settlement associated with thaw
consolidation amounts to 50–90% of heave. However, the two phenomena are
independent. For example, 30–80 mm of settlement occurred at Milchbruck
tunnel and was more or less the same for sections maintained by cyclic
refrigeration as by continuous cooling. Settlement can be controlled by grouting
during the thawing stage in order to fill any voids resulting from the freeze-thaw
cycle.
4
Drainage techniques
4.1
FILTER DRAINS
Drainage systems can be used to control high-pressure gradients or high pore-
water pressures (Cedergren, 1986). Such systems include aggregate filters and
synthetic filter materials, as well as pipes and other conduits. Most drainage
systems make use of porous filter aggregates to collect the water and conduct it
to outlets. Drains must be capable of removing all water that flows into them
without allowing an excessive build-up of head. They also should be designed in
such a way as to prevent the loss of fines from adjacent soil and, hence, avoid
becoming clogged. It is desirable to reduce the loss of head due to flow through a
filter to the lowest value compatible with the grain size requirements. Hence
effective aggregate filter drains almost invariably require one layer of graded
fine aggregate for filtering and a coarse layer of relatively high permeability to
remove ground water.
The following criterion has been used for the design of filters:
The ratio of D15 of the filter to D85 of the soil is termed the piping ratio. In other
words, the piping criterion dictates that the D15 of the filter material must not be
more than 4 or 5 times the D85 of the surrounding soil. If the piping criterion is
satisfied it is more or less impossible for piping to occur, even under extremely
la rge hydraulic gradients. According to the permeability criterion (on the right-
hand side) the D15 of the filter must be at least 4 or 5 times the D15 of the soil.
Although this means that filter layers are several times more permeable than the
surrounding soil it does not always guarantee hydraulic conductivity in drains.
Ideally the coefficient of permeability should be 10 to 100 times greater than that
of the average value of the surrounding soil. The piping criterion should always
be used to ensure against piping or clogging when pea gravel and coarse
aggregates are used in filter drains. In other words, the filter must be graded, that
is, the material should be enclosed by a layer of fine filter material.
Although the above filter criterion is frequently referred to as the Terzaghi
criterion, it was actually proposed by Bertram (1940). Since then there have been
many studies of filter criteria. For example, Anon. (1955) limited the piping ratio
to 5 and used the following criterion:
If filters are placed in plastic clay, then the US Army Engineers allow higher
piping ratios. In other words, for such soils the D15 size of the filter may be as
high as 0.4 mm and the above-mentioned criterion is disregarded. This relaxation
in criteria for protecting plastic clays allows the use of a one-stage filter material.
However, the filter must be well graded to ensure its non-segregation and therefore
a coefficient of uniformity (D60/D10) not exceeding 20 is necessary.
The results of tests carried out by Sherard et al. (1984a) on sand and gravel
filters showed that there was a very narrow boundary between filter failure and
success, when defined by
Uniformly sized particles of sand with D85 smaller than 0.1D15 (filter) always
passed through the voids of the filter, whereas when D85 (soil) exceeded 0.12D15
(filter) they were retained. In well-graded soils (e.g. sandy silts and clays, and
well-graded pervious sands) a filter sufficiently fine to catch the D85 (soil) size will
also catch the finer particles. It was found that for those soils tested, the most
widely used filter criterion
was conservative. It can be considered to have a safety factor of about 2. Thus

Sherard et al. (1984a) concluded that it is still appropriate to use
as the principal filter acceptance criterion. This conclusion applies generally to

filters with D15 larger than about 1.0 mm. For certain gap-graded and broadly
graded coarse soils, usually graded from clay sizes
DRAINAGE TECHNIQUES 69
Table 4.1 Sand or gravelly sand filters and fine-grained soils (from Sherard et al., 1984b)
Soil type D85 particle size of D15 particle size of Comments on D15
soil (mm) filter (mm) (filter) ≤5 D85 (soil)
criteria
Sandy silts and 0.1–0.5 Around 0.5 Satisfactory
clays
Fine-grained silts 0.03–0.1 Less than 0.3 Satisfactory
Clays with some Greater than 0.1 Around 0.5 Satisfactory
sand content
Fine-grained clays 0.03–0.1 Less than 0.5 Reasonable
Very fine-grained Less than 0.02 0.2 or smaller Satisfactory
clays and silts
to gravels, with D85 larger than 2 mm, the soil fines may be able to enter the voids
in the filter even if the coarser particles cannot. Sherard et al.’s (1984b)
recommendations for filters in silts and clays are summarized in Table 4.1.
It frequently has been suggested that a particle size distribution curve of the
filter should be approximately the same shape as that of the soil. Sherard et al.
(1984a) stated that generally this is neither necessary nor desirable).
Sherard et al. (1984a) also considered that the average particle size of a sand or
gravel filter, D50, does not provide a satisfactory measure of the minimum pore
sizes. Therefore, filter criteria using D50, such as
are unsatisfactory. They further maintained that filter criteria using D15 (soil), for
example,
were even less satisfactory. Hence, they recommended that the use of such filter
criteria should be abandoned.
One of the most recent surveys of filter criteria for cohesionless soils has been
undertaken by Honjo and Veneziano (1989). Their work showed that the stability
of the soil is controlled by the coarser particles, that is, a layer of coarser
particles forms at the soil-filter interface which, except for some gap-graded
soils, prevents the rest of the particles washing through. They confirmed that the
Terzaghi criterion was the most notable predictor of filter performance.
Nonetheless, they also found that a D95/D75 of 2 or less provides an indication of
the capability of the soil to form self-healing layers (i.e. to prevent grains
washing through). This they referred to as the self-healing index. They therefore
proposed a new criterion involving both the Terzaghi criterion and the self-
healing index, which was as follows:
Figure 4.1 Filter drain incorpoating synthetic fabric (geotextile), showing development of
a filter zone within the soil
Figure 4.2 Linear drains enclosed with geotextile
They maintained that this modified criterion provides a uniform safety (i.e.
probability of filter failure of about 0.1) over a wide range of soils. In addition,
they confirmed what Sherard et al. (1984a) maintained, that the ratio D50 (filter)/
D50 (soil) could not be justified on physical or statistical grounds and that the
Figure 4.3 Fin drain formed by a composite system of geotextile filter and plastic mesh
conducting core
requirement that the curve of grain size distribution of the filter should be
roughly parallel to that of the soil appeared to be unnecessary. However, their
single criterion for all soil types would, according to Talbot (1991), result in
filters which are too coarse for certain broadly graded soils and soils with
significant sand and gravel fractions. Hence, the filter criteria advanced by
Sherard et al. (1984a and b), which are based on the Terzaghi criteria, are more
reliable.
Vaughan and Soares (1982) showed that a relationship exists between the size
of particles retained by a filter and its permeability. Hence they suggested that
the permeability of a granular filter represented a better way of quantifying
particle retention than its grading.
The performance of a filter drain incorporating synthetic fabrics (Figure 4.1) is
governed, on the one hand, by the properties of the soil (particle size
Figure 4.4 Permeability reduction of geotextile installed in the field (after Van der Merwe
and Horak, 1989)
distribution, shape and packing) in which it is incorporated, and the hydraulic

flow conditions (undirectional steady state, transient laminar or turbulent flow,
dynamic/reversible flow, and angle of flow with respect to the plane of the fabric);
and, on the other, by the properties of the fabric (pore size distribution,
permeability, thickness, and variations of these with time and structural loading).
Fabrics have been used to form various types of drains such as linear drains
where fabric encapsulates granular material in a trench (Figure 4.2); drainage
blankets beneath fills; or fin drains (Figure 4.3). In the latter case the geotextile
consists of a conducting core enclosed between two layers of fabric (Anon,
1984). Van der Merwe and Horak (1989) have shown that the permeability of a
geotextile used as a filter does reduce with time (Figure 4.4).
When pipes are used in drainage systems, the slots in the pipes should be
small enough to retain soil particles thereby preventing their movement into the
pipes. Nonetheless, there should be enough slots to allow free movement of
water into the pipes. Obviously the filter materials around the pipes must have
gradations that are compatible with the sizes of the slots or holes. The following
criterion is frequently used
4.2
DRAINAGE OF SLOPES
Drainage is the most generally applicable method of improving the stability of
slopes or for the corrective treatment of slides, regardless of type, since it
reduces the effectiveness of one of the principal causes of instability, namely,
excess pore-water pressure. The distribution of groundwater within the slope
must be investigated, as must the most likely zone of failure so that the extent of
the groundmass which requires drainage treatment can be defined.
Surface run-off should not be allowed to flow unrestrained over a slope. This
is usually prevented by the installation of a drainage ditch at the top of an excavated
slope to collect the water draining from above. The ditch should be lined to
prevent erosion. It may be filled with cobble aggregate. Herringbone ditch
drainage is usually employed to convey water from the surface of a slope. These
drainage ditches lead into an interceptor drain at the foot of the slope
(Figure 4.5).
The interception of groundwater before it reaches an excavation is easier and
more effective than subsequent attempts to remove water once it has percolated
into suspect layers. Indeed, as much groundwater flow as possible should be
intercepted before excavation commences. For instance, deep interceptor drains,
placed at a high level on a site, prevent the development of high pore-water
pressures near the face of an excavation by stopping the penetration of water.
Water may be prevented from reaching a zone of potential instability by cut-
off, thereby allowing natural drainage to occur within the slope. Cut-offs may
take the form of a trench backfilled with asphalt or concrete, sheet piling, a grout
curtain or a well curtain, whereby water is pumped from a row of vertical wells.
Such barriers may be considered where there is a likelihood of internal erosion
of soft material taking place due to the increased flow of water attributable to
drainage measures.
Support and drainage may be afforded by counterfort drains, where an
excavation is made in sidelong ground, likely to undergo shallow parallel slides
(Figure 4.6). Deep trenches are cut into the slope, lined with filter fabrics, and
filled with granular filter material. Counterfort drains must extend beneath the
potential failure zone, otherwise they merely add unwelcome weight to the
slipping mass.
Horizontal drains have been used to stabilize clay slopes. They remove the
water by gravity flow. Consequently they have to be installed after the area has
been excavated a metre or so below the elevation at which the drains will outcrop
on the slope. The effectiveness of the drains is governed by their diameter and
spacing (the larger the diameter or the lower the spacing, the greater the increase
in slope stability), as well as their location in relation to a potential critical slip
zone. In other words, the amount of improvement depends on how closely the
drains are positioned in relation to the critical zone but there is no additional
Figure 4.5 Construction of surface drainage ditches filled with coarse granular material
leading into a lined interceptor drain which, in turn, leads into a sink. The material being
drained is till, near Loch Lomond, Scotland
benefit in extending the length of the drains beyond where this zone intersects
the top of the slope.
Successful use of subsurface drainage depends on tapping the source of water,
on the presence of permeable material which aids free drainage, on the location
of the drain on relatively unyielding material to ensure continuous operation
(flexible, PVC drains are now frequently used) and on the installation of a filter
to minimize silting in the drainage channel. Drainage galleries are costly to
construct and in slipped areas may experience caving. They should be backfilled
with stone to ensure their drainage capacity if they are likely to become partially
deformed by subsequent movements. Galleries are indispensable in the case of
large slipped masses (Figure 4.7), in some instances drainage has been carried out
over lengths of 200 m or more. Drill holes with perforated pipes are much
cheaper than galleries and are satisfactory over short lengths but it is more
difficult to intercept water-bearing layers with them. When individual benches
are drained by horizontal holes, the latter should lead into a properly graded
interceptor trench, which is lined with impermeable material.
Figure 4.6 Excavation A was made to provide a level recreational area for a school. The
ground, being near the bank of a large river in its lower reaches, consisted chiefly of
alluvium. Movement started on slip line SL but was unnoticed, for it was seasonal and not
very active. Later, a road was to be constructed for access to the school; this work
(excavation B) was found to be impossible until the hillside was stabilized. Counterfort
drains, C, were constructed about every 3 m over the length of the moving section. The
depth of the slip had been found to be about 2.5 m. It eventually became possible to build
the roadway
4.3
SAND DRAINS, SANDWICKS AND BAND DRAINS
The theory of consolidation advanced by Terzaghi (1925, 1943) proposed that
the time taken for a given soil to reach a certain degree of primary consolidation
varies directly with the square of the longest drainage path. In the case of
different types of soils the times taken are inversely proportional to permeability.
Obviously if the distance water has to travel in a mass of soil is reduced, then
consolidation will occur more quickly. This can be achieved by employing
drains.
A regular pattern of vertical drains permits a radial as well as a vertical flow
of water from the soil. The water from the drains is conveyed into a drainage
blanket placed at the surface or to highly permeable layers deeper in the soil.
With water escaping radially to the drains, in addition to escaping in the vertical
direction, the amount of time taken for consolidation can be reduced to a fraction
of that for vertical flow only, the time being governed mainly by the spacing of
the drains. Hence it can be concluded that the purpose of the drain is to provide
an easier path for the excess pore water to follow as it is squeezed out of a layer
of soil during consolidation. In addition to an acceleration in the rate of
consolidation, drains also bring about an acceleration in the rate of gain in shear
strength and reduction in the lateral transmission of excess pressure.
Sand drains are particularly efficient in stratified soils because of the higher
permeability parallel to the bedding (Figure 4.8). Indeed, sand drains have been
used in conjunction with dewatering systems to lower the groundwater level in
Figure 4.7 Internal drainage gallery in restored slope, Aberfan area, South Wales
stratified soils that contain alternate layers of pervious and relatively impervious
soil. Sand drains conduct water from the more permeable layers to the
dewatering screens which are located opposite the less permeable layers. This
reduces the cost of dewatering. However, it has been suggested that sand drains
probably serve no purpose where design loadings do not exceed preconsolidation
stresses. In such situations the rate of consolidation is more or less the same,
whether or not drains are used. Furthermore, drains may not be particularly
effective where secondary consolidation is the more important component of
settlement, as for example, in peats and organic clays. Nowadays wick or band
drains (see below) are more frequently used than sand drains as they are
generally more economic.
The coefficient of consolidation for radial flow and vertical compression
dominates the design of sand drains and is influenced primarily by the horizontal
permeability of the soil and to a somewhat lesser extent (in the sense that it is
less variable), by the coefficient of compressibility (Barren, 1948). It also varies
according to the type of loading process and the effective stress level, especially
around the preconsolidation pressure. The time, t, taken for consolidation brought
about by radial flow to a vertical well, can be obtained from
(4.1a)
Figure 4.8 (a) Derwent dam showing the complexity of glacial deposits under the deepest
part of the centre-line section; (b) Section through the dam and foundation, showing the
horizontal clay blanket linking the clay core with the clay cut-off, the vertical sand drains
which hastened the consolidation of the laminated clay, and the relief wells into the upper
aquifer (after Ruffles, 1965 Journal Institution Water Engineers 19)
(4.1b)
where Tr is the dimensionless time factor for radial consolidation, D is the

diameter of the equivalent cylindrical volume of soil being drained by the well, av
is the coefficient of compressibility (=∆e/∆p, i.e. change in void ratio with change
in loading), yw is the unit weight of water, kh is the horizontal permeability of the
soil, and e is the void ratio. The coefficient of consolidation for radial drainage in
a homogeneous soil can be determined from laboratory tests on undisturbed
samples and normally will equal or exceed values obtained from laboratory
consolidation tests with vertical drainage. Assessment of radial flow is
particularly important when sand drains are used in laminated soils (Rowe,
1964). However, design values for the coefficient of consolidation for radial flow
to sand drains must take into consideration the influence of variable permeability
within a formation, especially the influence of any continuous thin layers of sand
or silt that can dominate the permeability of a clay soil. The effect of such layers
can be estimated from the results of field permeability tests. Usually the ratio of
Figure 4.9 Effect of permeability ratio on sand drain efficiency (H=3m, D=3m, d =0.3m,
cv=0.096mm2/s) (after Richart, 1959, with permission of ASCE). H= height, D=area of
influence, d=diameter of drain, cv=coefficient of consolidation
horizontal to vertical permeability for such soils ranges between 5 and 10. The
influence of the permeability ratio on the effectiveness of sand drains is shown in
Figure 4.9.
The effects of soil disturbance due to the installation of sand drains that can
give rise to a smear zone about a drain should be taken into account when
assessing the coefficient of consolidation for radial flow (Richart, 1959). The
development of a smear zone depends upon how the hole for the sand drain is
formed in that mandrels are more likely to give rise to smear than augering. In
addition, the density and pore-water pressure of the soil may affect the
development of a smear zone. The extent of the smear zone may be quantified in
terms of the altered zone so that a smear ratio, SR, can be defined as
(4.2)
Normally when SR=1 the effect due to smear is of no consequence, whereas
when SR=1.2, then the drainage time can increase by up to 200%. Vertical
drainage probably will be as effective as sand drains if SR=1.5.
The spacing and pattern of drains are now fairly well standardized. Barron
(1948) developed curves (Figure 4.10) which related the average radial and
vertical excess pore-water pressures ur and uv, and average radial and vertical
percentage consolidation Ur and Uv with time factors for radial and vertical flow.
Curves are shown for values of n, that is, the ratio of the diameter of the zone of
influence, D, to the well diameter, d. In order to determine the well spacing that
corresponds to a given value of D, the vertical cylinder surrounding a well is
replaced by a vertical prism having an equal cross-sectional area. At the majority
of sites, triangular or square patterns with spacings between 1 to 4 m are used. It
Figure 4.10 Average consolidation rates (after Barron, 1948, with permission of ASCE).
The dashed line represents vertical flow in a compressible clay stratum of thickness 2H
drained at the top and bottom; the solid lines represent radial flow to axial drain wells in
clay cylinders having various values of n
has been shown that the diameter of the dewatered cylinder varies from 1.05
times the spacing when drains are positioned on a triangular grid to 1.13 times
the drain spacing when they are placed on a square grid (Figure 4.11). Spacings
of 1.5–2.5 m are the most commonly adopted.
The effectiveness of sand drains is influenced much more by their spacing
than by their diameter (Figure 4.12). For example, if the spacing between sand
drains is doubled, then this increases the time taken to achieve 90% consolidation
by roughly a factor of 6. On the other hand, by reducing the diameter of a sand
drain to a twentieth of its size, only increases the time taken for 90%
consolidation by a factor of approximately 4. In fact the time for consolidation is
proportional to the 2.5 power of the drain spacing.
The depth to which drains are installed is normally equal to the thickness of
the soil concerned. For depths up to 20 m vertical drains often prove an
economic solution. Beyond 20 m depth, however, the cost of placing drains
increases sharply due to the extra effort involved in their installation. For
instance, even normally consolidated clays are firm to stiff at depths of 30–40 m.
In fact, McGown and Hughes (1981) questioned the value of treating such depths
since most settlement and likelihood of shear failure occurs at shallow depths.
Sand drains generally are constructed by inserting casing into the ground and
then placing sand in the hole, under air pressure as the casing is withdrawn. The
casing may be driven or jet placed and the soil displaced by using a closed-end
mandrel, jetted from the hole (Figure 4.13), or removed by augers or rotary-
drilling methods.
The installation of sand drains means that consolidation occurs at a faster rate
near the sand drains, thus causing a larger amount of surface settlement in that
region. Such differential settlement can be responsible for a redistribution of
surface loading. In fact sand drains of large diameter in a foundation may act as
piles in soft soils, attracting load and somewhat relieving stress in the
surrounding compressible deposit.
Figure 4.11 (a) Typical vertical drain installation layout; (b) theoretical pattern of drainage
through a sand drain
Generally, a drainage blanket of granular material is laid over the area where
vertical drains are installed. A considerable quantity of water may be discharged
into the drainage blanket, especially during the early stages of construction, if a
large volume of soil is being drained. The amount of water that can be
discharged into the drainage blanket from a given layout of drains can be derived
for any particular degree of consolidation. In this way the effectiveness of the
drainage blanket can be assessed and, if necessary, its design can be adjusted
accordingly.
As a reduction in the diameter of a sand drain reduces the efficiency of
drainage by a slight extent, and as it has proved difficult to construct sand drains
with diameters smaller than 250 mm, the sandwick was introduced (Dastidor et al.,
1969). Sandwicks are formed by pneumatically filling stockings made of woven-
bonded polypropylene with graded sand. The most convenient diameter for
sandwicks is around 65 mm. They are installed in holes of small diameter and
have proved effective when used to consolidate suspect alluvial deposits.
Figure 4.12 Effect of drain installation and drain spacing and diameter on consolidation
time for a clay soil (in (a) d=0.3 m; in both kh=5kv and cv=0.096 mm2/s) (after Richart,
1959, with permission of ASCE)
Sandwicks can be emplaced by a variety of methods depending on the type of
soil involved (Robinson and Eivemark, 1985). These include solid or hollow
stem flight auger rotary drilling, rotary wash boring, jetting, and driven and
vibrated casing. They can be placed to greater depths than sand drains and the
fabric stocking ensures continuity at the time of placement.
Although a larger number of sandwicks are needed to bring about the same
rate of consolidation as small sand drains, sandwicks have several advantages.
The fabric stocking means that a sandwick is continuous, and flexible, and so can
adapt to vertical settlement or lateral deformation. It acts as a filter so reducing
the possibility of clogging. The likelihood of intercepting lenses of dubious
material in heterogeneous soils is far greater if sandwicks with smaller spacings
are installed rather than more widely spaced sand drains.
The efficiency of a drain largely depends on its cross-sectional circumference
rather than on its cross-sectional area. Hence band-shaped drains are more
effective than drains having a circular crosssection. Since the early 1970s man-
made fabrics—notably polyethylene, PVC, polypropylene and polyesters—have
been used increasingly to produce band drains.
Figure 4.13 Emplacement of a sand drain. Equipment is designed to jet through varying
subsoils and leave a vertical column of graded media in the ground
Several types of flat band drains are available (Table 4.2). Most consist of a
flat plastic core containing drainage channels surrounded by a thin filter layer
(Figure 4.14). Engineering filter fabrics are now more commonly used than
treated paper wrappings. Lee et al. (1989) referred to the use of fibre drains made
from jute and coir which have functioned for periods of 12–15 months. In
countries where these materials are abundant, the principal advantage of fibre
drains is their low cost. Although the permeabilities of the wrappings, especially
of paper, may be as low as 10−8 m/s, they are thin and offer little resistance to
flow. The permeability of the filter should not be higher than that required for the
discharge capacity (Hansbo, 1979). This means that filter permeability of
existing prefabricated drains need not exceed 0.01–0.1 m/y, irrespective of
whether drains are placed in clay or silt. The mechanical characteristics of band
drains are important. In particular, the tensile strength usually ranges from 0.5 to
2 kN and elongation at yielding is between 6 and 10%.
The discharge capacity of band-shaped drains varies considerably, depending
upon the type of drain used. It also is a function of the effective lateral earth
Figure 4.14 Typical cross-sections of plastic band drains: (a) with filter sleeve; (b)
without filter sleeve
pressure against the drain sleeve. In most cases the filter is partially squeezed
into the channel system of the core due to the pressure of the surrounding soil.
Accordingly, the discharge is lowered as a result of the reduction in cross-
sectional area. Furthermore, the discharge capacity is reduced by ‘ageing’ once
the band drain is installed (Figure 4.15). Other factors that may reduce the
effectiveness of band drains include fines entering the channel system and
buckling of the drain under large vertical strains. According to Bergado et al.
(1991), smear effects due to installation may significantly reduce the
effectiveness of band drains. If a drain functions correctly during the very early
stages of consolidation, the excess pore water contains a small amount of soil
particles which may clog the drain. Those drains that do not have open channels
in their central area are especially susceptible.
Band drains generally are installed by displacement methods, usually
Table 4.2 Main characteristics of some common types of band drains

Data Units Geodrain Geodrain KB Desol
Drain body material Polyethylene Polyethylene LD Polyolefine
Filter sleeve material Cellulose fibre Polyester Cellulose Without
Grooves (number) 27×2 27×2 24
Weight (including sleeve) g/m 168 152 44
Width mm 95 95 95
Thickness mm 4 4 2
Free surface area mm2 220 220 150
Data Units Geodrain Geodrain KB Desol

Lateral water permeability m/s 10−6 10−6 10−3
Length in each roll m 150 150 400
Roll weight kg 28 28 20
40 ft container m 59 400 59 400 160 000
by a lance (Figure 4.16). Speeds of installation are generally of the order of 0.3–
0.6 m/s. Auger and rotary wash-boring methods usually are unsuitable, but water
jetting within an open casing may be used in very soft soils. Experience indicates
that drains spaced at 1–2.5 m centres and extending to depths of between 10 and
20 m give the highest production rates for most methods of installation. In
practice, drain spacing is rarely
Figure 4.15 Results of discharge capacity tests carried out on geodrains that have been
pulled up after having been in peat (left) and slime (right) for different lengths of time
(number of days in soil given in figure). Results obtained with filter sleeves of paper
shown with full lines and with filter sleeves of synthetics with broken lines (after Hansbo,
1987)
Table 4.2 con’d
Alidrain Mebradrain MD 7007 Colbond CX 1000 PVC

Polyethylene 100% polypropylene Coarse polyester monofilament PVC
Polyester Polypropylene Polyester Without
32×2 38 48 14
180 92 110 100
100 100 100 (150–300 possible) 100
Alidrain Mebradrain MD 7007 Colbond CX 1000 PVC

5 3 5 1.8
180 180 220 56
10−5 10−3 10−3 10−5
100–150 250 200 800
23 25 90
110 000 80 000
Figure 4.16 Placement of band drains

less than 0.8 m. Band drains have been installed to depths of 40–60 m (Hansbo,
1987).
The effects of smear caused by installing band drains by displacement
methods can be reduced by the correct choice of drain filter fabric and the size of
the installation lance. A filter fabric initially allows the finer particles of soil,
notably those of clay size, to pass through. Hence piping occurs, removing fines
from the smear zone, which leads to the formation of a natural graded filter in the
soil that may be several millimetres in thickness. This only occurs when the drain
fabric has appropriate sizes to allow the development of a natural graded soil
filter. Clogging occurs if the pores are too large whereas if they are too small the
effect on the smear zone is limited.
4.4
LIME COLUMNS
Cylindrical columns can be formed in clay soils by mixing the clay with
unslaked lime (see Chapter 10). This increases the permeability of the columns to
between 100 and 1000 times greater than that of the surrounding clay (Broms
and Boman, 1979). Hence the lime columns act as vertical drains, as well as
reinforcing the soil. One lime column of 500 mm diameter has the same drainage
capacity as three 100 mm wide band drains. One of the advantages of lime
columns is that their installation creates little disturbance in the surrounding soil,
which enhances their performance as a drain.
5
Groundwater lowering
Dewatering is one type of temporary measure taken to lower the groundwater

level so that excavation can take place in dry, and therefore more stable,
conditions (Bell and Cashman, 1986). However, a change in the position of the
water table brought about by drawdown, particularly around a deep, wide
excavation may adversely affect the area surrounding the excavation. Of special
concern is any settlement likely to occur as a result of withdrawal of
groundwater.
5.1
SUMPS
Dewatering brought about by pumping from sumps is the simplest, most widely
used—and in many cases the most economical—method of lowering the
groundwater level. Pumping from a sump is often the first method tried where an
unexpected problem with groundwater is encountered. However, because each
sump requires its own pump, the method becomes impractical when more than a
few sumps are necessary. A sump must be deep enough to ensure that the entire
excavation is drained as a result of pumping. The open sump method is only
capable of lowering the water table by up to approximately 8 m. The sump(s) are
located below the general base level of the excavation at one or more corners or
sides. Continuous pumping of groundwater can remove fines from the
surrounding soil if the sump is not surrounded by an adequate graded filter. A
ditch or garland drain, which falls towards the sump, is dug around the base of the
excavation to convey water from the slopes to the sump. The ditch should be
lined or filled with gravel to prevent erosion taking place. Perforated pipe can be
placed within the gravel to increase the rate of water removal. Lining the ditch
with a filter fabric helps retard the migration of fines.
The pumping capacity needed to establish the initial lowering of the water
table is about twice the steady state of equilibrium rate of pumping. Therefore, if
double the required steady rate capacity is installed, then all pumps can be used
in the initial drawdown period. Subsequently, half the pumps can be shut down
and used for emergency standby and maintenance purposes.
Figure 5.1 Wellpoint system dewatering an excavation

5.2
WELLPOINT SYSTEMS
A wellpoint system consists of a number of small wells installed at close centres
around an excavation (Figure 5.1). An individual wellpoint usually consists of a
gauze screen 0.5–1.2 m long and from 40 to 75 mm in diameter, which surrounds
a central jetting riser pipe (Figure 5.2(a)). Various sizes of screen openings are
available to correspond with the predominant grain size of the soil. Cashman
(1975) recommended that where wellpoints are required to remain in the ground
for more than a few weeks, it is more economical to use disposable plastic
wellpoints. These consist of a nylon mesh screen surrounding a flexible plastic
riser pipe (Figure 5.2(b)). Alternatively, PVC wellpoints, which consist of a
plugged tube with an array of fine slots cut through its bottom section, can be
used (Figure 5.2(c)). An inner tube is arranged so that the water draining into the
wellpoint is drawn upwards from just above the base plug. In both cases the
wellpoint and riser are simply cut to various lengths to suit ground conditions
and depth. At the end of the dewatering period the riser pipe is cut off at or below
the ground surface. In fact plastic pipes now tend to be used instead of metal
pipes. The individual wellpoints are connected by vertical riser pipes to a header
main which is under vacuum from a pumping unit (Figure 5.3). The water flows
GROUNDWATER LOWERING 89
Figure 5.2 (a) A jetting shoe. The latter is attached to the riser pipe. A strainer is located
in the recess of the jetting shoe and retained by the retaining collar; (b) Disposable
wellpoint and riser pipe; (c) PVC wellpoint
under gravity to the wellpoint and is drawn by the vacuum to the header main
and discharged through the pump.
Because the pumps are on the surface and draw air through the wellpoints,
once the groundwater level is lowered they generally have to pass more air than
water. This situation can be dealt with by doubling the air capacity of the pumps
by using twin-diaphragm attachments instead of a single-diaphragm self-primer
(Anon., 1989(a)).
A cone of depression is developed about each wellpoint. The individual cones
should intersect in such a way as to lower the water table 0.5–1.0 m below the
base of the excavation concerned. As the wellpointing system draws the water
away from the excavation, this means that its slopes are stabilized and therefore
can be steeper.
The system is most suitably employed when the water table does not need to
be lowered more than about 6 m. Multi-stage systems generally are required
beyond this depth because of the limitation of suction lift (Figure 5.4).
Nonetheless, improvements in the system have meant that higher vacuums can
Figure 5.3 Wellpoint system showing wellpoint, riser pipe, header main and pump
Figure 5.4 Multi-stage wellpoint system

be maintained, so achieving higher suction lifts and enhancing the ability to
dewater fine-grained soils (Figure 5.5).
The main advantage of a wellpoint system is its versatility (Cashman and Haws,
1970). In particular, it can be readily adapted to unexpected groundwater
conditions: for instance, wellpoints can be added at offending locations,
additional pump capacity can be supplied and wellpoint screens can be raised or
lowered with a minimum of trouble.
Figure 5.5 Soil types and dewatering systems
The most commonly used wellpoint is the self-jetting type, which can be
installed rapidly (often in a matter of a few minutes) by jetting water from the
nozzle of the wellpoint (Werblin, 1960). On average, around 50 wellpoints can
be installed in one day. However, under ideal conditions in clean sand, up to 100
may be put in place. Conversely, with difficult jetting conditions output may
drop significantly. Each wellpoint needs about 1 m3 of water or more to jet it into
the ground. The water frequently has to be supplied from a tanker, but once a few
wells have been established ground and jetting water can be recycled into a tank
on site. Self-jetting wellpoints are suitable for installation in gravels and sands,
and can penetrate silts and soft to firm clays. They are suitable for soils that do
not yield more than 40–100 1/min, depending on diameter and screen area. For
capacities up to 140 1/min per wellpoint, highcapacity wellpoints are required but
they are not jetted into place. Large-diameter wellpoints (160–200 mm) are
referred to as suction wells and are used where the capacity exceeds 140 1/min.
Occasionally problems arise when jetting occurs in soils containing highly
permeable zones of very coarse sand and gravel. The operation can be suspended
if the jetting water disappears into such zones. In such cases the wellpoint should
be raised a metre or so above the offending zone and held there until turbulence
reappears. It then should be lowered slowly towards the coarse layer.
Various drilling methods have been used to install wellpoints when jetting is
slow due to the soil conditions. For example, drilling with augers, followed by
either plain jetting or installation within a casing is common. If a bed of clay
overlies the formation to be dewatered, then it is usually more practical to auger
than jet through the clay. A hole-puncher may be used to install wellpoints in
soils which offer considerable resistance to penetration such as coarse gravels,
cobbles or stiff clays, as well as very permeable soils which are subject to loss of
boil (Figure 5.6). After reaching the desired depth the cap is removed and the
Figure 5.6 Hole-puncher and casing
wellpoint installed before the hole-puncher is extracted. When equipped with a

sanding casing, the hole-puncher has been used to jet holes up to 600 mm in
diameter and 36 m in depth.
A filter surrounds the wellpoint and riser and increases the effective diameter
of the wellpoint, which increases the amount of discharge. The filter column
above the screen provides vertical drainage through any thin layers of silt or clay
present in sands or gravels. In highly permeable gravels a wellpoint usually does
not require a sand filter around it, but if it does, then the wellpoint must be
positioned in a cased borehole. The casing is withdrawn after the filter sand is
placed.
Testing and evaluating the performance of a dewatering system as it
is installed and operated are advisable because, owing to normal variations in
soil conditions, it is usually impossible during the design stage to determine with
accuracy the permeability of strata, the radius of influence or distance to the
effective source of seepage, and losses of head in the wellpoint system.
The spacing of wellpoints about an excavation is governed by the permeability
of the soil and the time available to effect drawdown. Where the aquifer extends
3 m or more beneath the subgrade, the selection of wellpoint type and spacing is
based on the quantity of water to be removed. The total flow of the system is
estimated and divided by the length of the header main to obtain the flow per
unit length. The wellpoint spacing is then chosen.
Figure 5.7 A nomogram to determine spacing between wellpoints
Normally wellpoints are spaced between 1 and 4 m apart (Figure 5.7). Wells
probably should be considered if the spacing is greater than 5 m. Dewatering an
unconfined bed of sand or sandy gravel of moderate permeability above an
impermeable bed represents one of the best situations for the employment of
wellpoints (Mansur and Kaufman, 1962). Under special conditions, however,
narrower spacings may be required. For instance, thin layers of clay often occur
in water-bearing sands, and because they are impermeable they give rise to
interruptions in the drawdown curves around wellpoints. In such situations
wellpoints commonly are spaced 1 or 2 m apart. Alternatively, holes can be
positioned outside the line of wellpoints and filled with coarse sand. Such
columns of sand provide a vertical path in the zone of sand being drained and
enable water to move more readily towards the wellpoints than towards the sides
of the excavation.
Figure 5.8 A concentric dewatering system using wellpoints and deep wells
Conversely, difficulties can also be brought about by layers of highly

permeable material. In this case water tends to by-pass the wellpoints. The
problem can be overcome by jetting at close intervals in a line around the
excavation, thereby mixing the soils of the different layers.
Wellpoints using vacuum-assisted drainage are effective in silty sands and
silts, their spacing usually ranging between 1 and 2 m. A plug of bentonite or
cement is placed at the top of the filter column to seal against the entrance of air.
Vacuum is applied through the filter column to the soil, thereby accelerating
drainage.
In practice the suction lift of a single-stage wellpoint is limited to about 5 m.
Under ideal conditions, and using high-vacuum equipment, the depth of lowering
can be increased to around 7.5 m. If the total drawdown required is appreciably
more than 7m, it is generally necessary to use a multi-stage wellpoint system, or
a combination of deep wells and a single stage of wellpoints or an ejector system
(Figure 5.8). On average each extra stage or tier can lower the water table some
4.5 m.
There is no theoretical limit to the depth of drawdown if a multi-stage system
of wellpoints is used. However, the overall width of the excavation has to be
increased in order to accommodate this type of layout. Sometimes the upper tiers
can be reduced in number or removed after the lower ones have been brought
into operation. If an excavation is taken down to the water table before the
header main and a pump is installed, then fewer tiers may be used.
The average thickness of the outer part of the slope drained by a multi-stage
system is not more than about 4.5 m. Beneath this drained layer, the soil is
subjected to seepage pressures produced by percolating water. The stability of
the slope should be assessed if the depth of the excavation is more than 12–15 m.
This assessment should consider the seepage forces beneath the drained zone.
Figure 5.9 Installation of horizontal wellpoint by purpose-built machine (from

Somerville, 1986 CIRIA Report 113)
Where the depth by which the water table has to be lowered is more than 4.5 m
but the rate of pumping from each wellpoint is relatively small (that is, less than
45–65 1/min), then the installation of a single-stage system of wellpoints at the
top of the excavation or water table, each wellpoint having an ejector pump, may
prove more successful than a multi-stage system of wellpoints.
Once the wellpoint system has been established, it is then necessary to dispose
of the water. If the system is working properly, it should not be drawing water
laden with silt. Most of the fines come through the system in the initial 20
minutes after the start of pumping, and within about 4 hours the water should be
clear. Fines will only be drawn off again if pumps are closed down and then
restarted.
Horizontal ‘wellpoint’ drainage is particularly suited to ground water lowering
for pipe trenching where fast forward progression is required (Figure 5.9). A
long perforated pipe is laid horizontally in the soil at the desired depth by a
special trenching machine. This excavates a trench 225 mm wide at depths
between 1.8 and 6.0 m and lays thin-walled convoluted perforated plastic pipe of
80–100 mm diameter in the bottom of the trench as the trencher tracks forward.
The pipe is wrapped with a mesh of woven nylon or coco-matting. The machine
throws the dug spoil back into the excavated trench on top of the perforated pipe.
The perforated pipe is cut at given lengths and connected to plain pipe which is
brought to the surface at the end of each run. This pipe is then connected to a
wellpoint pump. The rapidity with which this system can be installed means that
the contractor can progress speedily. This is particularly relevant when laying
natural gas pipelines in open country.
5.3
EJECTOR OR EDUCTOR SYSTEMS
An ejector is a jet pump that is used for raising water from a borehole. According
to Miller (1988) it is especially well adapted to operating as a group to form a
dewatering system. Ejector pumps remain operational in a borehole whether or
not they are drawing water. However, they have limited flow capacity so that
they are restricted to dewatering less permeable soils.
The simplest type of ejector consists of a downpipe which carries pumped
water to an upward-facing nozzle, from which a high velocity jet of water enters
the venturi section (Figure 5.10). This jet induces ground water to enter the
venturi and both ground and pumped water flow up the riser pipe to the surface
into a tank (Powrie and Roberts, 1990). The induced flow of groundwater is
brought about by momentum transfer in the highly turbulent perimeter of the jet
(Miller, 1988). Nozzle velocities of around 30 m/s are required to lift water from
20 m depth and 50 m/s to lift from 50 m. Water needed to maintain the operation
is recirculated from the tank and the excess disposed of. Cup seals have to be
incorporated beneath a single pipe ejector to provide an end-stop to the annular
downpipe, thereby preventing pumped flow discharging into the ground.
If the groundwater level is at or below the intake level of the ejector, then air
flowing into the throat of the ejector develops a vacuum inside the well,
accelerating drainage and reducing pore-water pressures. When air is not
available, subatmospheric pressure conditions can develop. This subatmospheric
pressure can trigger chemical precipitation, which can cause scaling problems.
For example, when ground-water contains more than 50 mg/1 of iron its
precipitation can block ejectors. Accordingly, an ejector may have to be taken out
of service and cleaned occasionally.
Ejector systems are most effective when used to dewater deep excavations
made in stratified soils where close spacing is necessary (Figure 5.11). In an
ejector system the pumped flow inlets to the individual ejectors are connected to
a common pumped flow main. Similarly, the return flow outlets are joined to a
common return flow main which discharges into a water storage tank. Junctions
between mains and individual ejector well heads have shut-off valves and a
union connector so that any ejector can be readily isolated from the rest of the
system. Ejector systems are not limited in terms of suction lift as are wellpoints,
and have much lower unit costs than wells. Hence, they are used where it is
inconvenient or impossible to install a multi-stage wellpoint system since the
principal advantage of the ejector system is its ability to operate numerous
wellpoints in a single tier to depths in excess of 20 m. Individual ejectors can
pump from 20 to 270 1/min (Prugh, 1960). They are installed by jetting within a
casing in a similar way to that used for sinking conventional wellpoints in difficult
Figure 5.10 (a) Single-pipe ejector; (b) two-pipe ejector

ground, and are surrounded by a sand filter. They can also be installed by water
flush rotary drilling, and surrounded by a perforated liner and sand filter.
Spacings of 3 m to 15 m are typical. According to Miller (1988), it is not
normally economical to space a line of ejectors much closer than 3 m.
The ejector is a self-priming device. It evacuates air from the wellpoint
beneath. If one ejector in a system is drawing air, it does not affect the balance of
the system, provided air can be vented from the return or collector main. When
the flow of ground water is less than the design capacity of the ejector and the
filter column has been sealed properly, then the unit develops a vacuum in the
Figure 5.11 An ejector or eductor system
wellscreen and this is transmitted to the soil. The ability to develop a vacuum in
the soil is particularly effective in draining fine-grained soils. Indeed, the
maximum practical capacity of an ejector-operated system is 6800–9000 1/min,
which usually limits their application to fine sands and silts. If the wells are
carefully constructed using appropriate filters and bentonite seals, the effects on
fine sands and silts can be dramatic. Only low volumes of water have to be
removed from such soils to achieve stabilization so that the low efficiency of
ejectors is acceptable.
5.4
BORED WELLS
A deep bored well system consists of three components, namely, the well itself,
the pump and the discharge piping (Figure 5.12). The well has an open area into
which groundwater flows and a cased section which conducts the water to the
surface. In granular soils, slotted metal screens or perforated casings are used to
prevent collapse of the walls in the well zone. A graded filter is placed around
the screen. During the drilling operation a temporary outer casing is inserted in
the drillhole, inside which the permanent inner casing and screen are positioned.
The annular space between the two casings is then filled with the filter.
Deep wells have been used for lowering the water table where the soil
conditions become more pervious with depth, and have proved particularly worth
while where it has been necessary to remove large quantities of water from thick,
highly permeable formations. Indeed, deep wells are most effective where the
water-bearing soil is clean, uniform and highly permeable such as some sands
and gravels. They also are effective where a large excavation is underlain by a
relatively impermeable bed, beneath which there is a permeable formation under
excessive artesian pressure.
For dewatering purposes deep wells commonly are located at 6.5–65 m
centres, depending upon the quantity of water that must be removed, the
permeability of the ground, the source of seepage and the amount of
submergence available. Such wells usually have a diameter of 150–450 mm,
although at times the diameter may exceed 1 m.
The hydraulic efficiency of a deep-well system may be increased by some 10–
15% by applying a vacuum line connected to the tops of the wells. The wellhead
and annulus must be airtight.
A deep-well system is installed by boring or drilling operations. The selection
will depend upon the soil conditions expected and the availability of equipment.
Holes extending to 30 m in depth with diameters of up to 600 mm have been
excavated by bucket augers. Bucket augers are effective in gravels and sands,
and in soft to moderately stiff silts and clays. However, cobbles and boulders
present problems and very stiff clays are difficult to penetrate. Excavation by
bucket auger can yield a hole of good quality if caving is prevented by using a
head of water. A minimum of about 3 m of head is recommended (Powers,
1981). Subsequently, the hole requires little development. In loose permeable
sands it may be necessary to mix additive with the boring fluid to form a
temporary seal around the hole, thereby preventing caving. Bentonite is not
recommended because the filter cake it forms is difficult to remove from the
hole, but biodegradable muds are acceptable.
Dewatering wells are occasionally constructed by continuous flight augers.
Heavy duty augers can excavate holes with diameters up to 1.2 m. At times
heavy slurry is required to stabilize the hole and to aid the retention of sand on
the auger as it is being removed. The slurry, together with the mixing of soil and
Figure 5.12 (a) A deep-well system; (b) well in more detail
the smearing of the sides of the holes, usually results in a hole of poor quality.
Rotary drills, where the cuttings are removed by circulating fluid, are effective
for holes of small to moderate diameter. For example, a heavy duty rig can
produce a hole 450 mm in diameter.
Reverse circulation rotary drilling is sometimes employed in dewatering
programmes. Holes with diameters of 600 mm or more can be sunk. This type of
drilling generally results in a cleaner hole than the conventional rotary methods.
The method is best suited to loose sands and gravels and soft clays. Stiff clays
prove difficult to penetrate and cobbles may also cause difficulty since they tend
to become lodged in the drill pipe. A head of water is required to provide support
for the sides of a hole when using reverse circulation drilling, a minimum of 3 m
from the drilling surface to the water table being recommended. Where the water
table is close to ground level the drilling rig can be placed on a platform or,
alternatively, a small wellpoint system can be operated while the hole is being
bored.
The development of a well is an essential operation in its proper completion.
There are three beneficial effects of development:
1. Development corrects any clogging of the water-bearing formation that may

have occurred during drilling, and it effects removal of boring slurry.
2. Development tends to increase the porosity of the natural formation in the
immediate vicinity of the well and so improves its yield.
3. Development stabilizes the filter media formation stabilizer in the annulus.
In addition, removal of ‘fines’ during development results in clean water being

pumped when the electro-submersible pump is installed (i.e. no fines in
suspension), thereby prolonging the useful life of the pump.
There are two basic techniques of well development, namely, surge plunger,
either solid or valve type, and air-lift pumping. In general the air-lift pumping
technique is to be preferred although it is not effective for relatively shallow
wells, that is, less than about 20 m deep.
Depending upon ground conditions, deep wells either may be used exclusively
to dewater an excavation or they may be used in conjunction with another
pumping system. For instance, wellpoints may be required at the base of the
slopes of an excavation in order to intercept any minor seepage that may by-pass
the wells. Alternatively, heave or a blow-out may be a possibility where the
strata beneath the floor of an excavation are subject to artesian pressure. The
artesian pressure can be relieved by the installation of deep wells either around
the top of the excavation penetrating into the offending aquifer or at some safe
elevation as the excavation is extended downwards. Depending upon the
magnitude of the excess pore-water pressure and the soil conditions, relief wells
or ‘bleeder’ wells may be adequate.
Where a site consists of silts or silty sands that are underlain by a bed with
higher permeability—for example, of coarse sand—seepage towards the
excavation can be intercepted and the water table lowered by a combination of
vertical sand drains installed around the top of the excavation and deep wells
installed in the coarse sand. The sand drains facilitate drainage of the upper, less
permeable soil, while the pore-water pressure in the coarse sand is reduced by
pumping from the deep wells (Powers, 1981).
Shallow-bored filter wells are a synthesis of wellpointing and deep wells—
some refer to this method as ‘jumbo’ wellpoints. The installation of the wells is
the same as for deep-bored wells with the same facility to ensure controlled
filtering, but the wells are pumped by suction pumping either by individual risers
to a common suction main or by individual wellpoint pumps at each wellhead.
On a congested site, especially in high-permeability soils, this method may be
preferred to wellpoints because of the smaller number of risers causing
obstruction to the construction operation.
5.5
SETTLEMENT AND GROUNDWATER ABSTRACTION
When a water table has to be lowered in order to construct an excavation, the
effect of such an operation on neighbouring structures must be considered. On
occasions it may be necessary to keep the groundwater at its original level in a
built-up area immediately adjacent to an excavation so that the existing
structures are protected from possible damage.
Lowering the water table by abstracting groundwater leads to an increase in
effective pressure, which, in turn, can give rise to settlement of the ground
surface. Significant settlements can also occur when water is pumped from a
confined aquifer overlain by compressible soil. Similarly, abstraction of water
from an aquifer containing layers or lenses of compressible soils may cause
settlement. The amount of settlement that occurs depends on the thickness of the
compressible soils and their compressibility, on the amount by which the
phreatic surface is lowered and on the length of time over which pumping takes
place (Placzek, 1989). The rate of settlement is governed by the permeability of
the compressible soil.
Depending on the site conditions, it may be feasible and desirable to use
groundwater recharge to limit the radius of influence of the cone of depression
and so reduce potential settlement. In such instances the excavation usually is
surrounded by sheet piling with the recharge wells located outside the piling
(Figure 5.13). Relief wells within an excavation can be used to reduce
piezometric pressures, and thereby overcome the problem of heave.
5.6
ELECTRO-OSMOSIS AND ELECTROCHEMICAL
STABILIZATION
Elecro-osmosis was originally developed as a means of dewatering fine-grained
soils (Casagrande, 1952; Lo et al., 1991). It consists of passing a direct current
from anodes to cathodes positioned at predetermined locations in the soil to be
Figure 5.13 A pumping and recharge system to avoid settlement during the excavation of
basements (after Zeevaert, 1957)
stabilized (Figure 5.14). As the current passes through the soil it causes water to
migrate from the anodes towards the cathodes where it is collected and then
removed (Figure 5.15). The mechanism responsible for bringing about migration
is not completely understood but it appears that absorbed water is removed from
clay particles, and that ion transfer through the pore water causes it to flow.
Not only does electro-osmosis reduce the water content of the soil but it also
directs seepage forces away from the surface of an excavation as well as
developing pore-water tension in the soil (Perry, 1963). Moreover, base
exchange occurs in the soil during electro-osmosis, thereby enhancing its
strength. In addition, ion migration, electrolysis and chemical reactions occur
leading to the formation of new irreversible compounds. Indeed, chemicals such
as calcium chloride and sodium silicate are sometimes added to ensure the
growth of cementitious material in the pore space.
Figure 5.14 Layout for dewatering by electro-osmosis
Figure 5.15 A typical cathode well system
Where electro-osmosis has proved successful, it usually has been as a result of

using the process to introduce a chemical into the soil, either through anode
solution or by direct electrolyte replacement. This improves the stability of the
soil either by ionic replacement occurring in the clay mineral content, or by
cementitious material being deposited in the pore space. This adaption of electro-
osmosis is known as electrochemical stabilization.
Unlike the coefficient of soil permeability (k), and therefore the rate of
hydraulic (laminar) flow, the coefficient of electro-osmotic permeability (ke) and
the rate of electro-osmotic flow are independent of the size of the capillaries.
Hence the electro-osmotic permeability is more or less independent of the soil
pore size. In other words, it is about the same magnitude for sands, silts and
clays, provided that the electro-osmotic potential is about the same for most of
the mineral matter in the soil. The electrokinetic potential depends on the
concentration and valence of soluble ions in the electrolyte. It is high in a diluted
solution, whereas in a concentrated solution it may be reduced to zero. In fact,
the potential does not vary much within the ordinary range of concentration of
electrolytes present in soils. However, the electrokinetic potential can fall to zero
in some situations and the electro-osmotic flow either ceases or may be reversed
in direction. For practical purposes, most saturated soils may be assumed to have
a coefficient of electro-osmotic permeability of around 2×10 to 5×10 cm/s for a
gradient of 1 volt/cm—for example, Norwegian quick clay=2×10 cm/s; clayey
silt=5×10 cm/s; and fine sand=4.1×10.5 cm/s. Electro-osmosis can be used for
dewatering silts and clays but the method is not suitable for deposits of sand.
The relative efficiency of electro-osmosis as a dewatering technique may be
expressed in terms of the ratio between ke and k referred to as the coefficient of
electro-osmotic effectiveness. It increases with increasing clay fraction in the
soil.
Spacing of electrodes depends on the electrical potential available, the
potential gradient required for a given soil and ground water conditions.
Spacings between 3.6 and 4.95 m and potentials varying between 30 and 180
volts have proved effective. As far as individual soil types are concerned Zhinkin
(1966) proposed a potential gradient of 0.9 volt/cm for sandy loams, 0.7 volt/cm
for loams, 0.6 volt/cm for clays, and 0.4 volt/ cm for silts. This implies that for a
direct current potential of 100 volts, the electrode separation would range from
just over 1 m to 2.5 m. For maximum effectiveness, the potential gradient should
be in the same direction as the hydraulic gradient. Generally, potential gradients
of more than 0.5 volt/cm should not be exceeded for long-term applications
because higher gradients lead to energy losses in the form of considerable
heating of the ground. However, it might prove advantageous to use potential
gradients of 1–2 volt/cm during the first few hours. This gives a much faster
build-up of tension in the pore water. Economy of power consumption could
perhaps be obtained by intermittent operation.
Negative pore-water pressures develop as water is removed next to the anodes
while at the cathodes the pore-water pressures can increase to high positive
values. This difference in pore-water pressure produces a hydraulic backflow. It
also means that the soil around the cathodes may be softened.
Consolidation may occur as a result of the development of negative pore-
water pressure under controlled electro-osmotic drainage (Mitchell and Wan,
1977). This, in turn, leads to changes in the stress-strain and strength
characteristics of the soils concerned. The Terzaghi equation of consolidation for
homogeneous material also applies to electro-osmotic consolidation where the
voltage distribution remains constant with time. Like ordinary consolidation
under direct loading, electro-osmotic consolidation is a mass flow process in
response to pressure of potential gradients. The rate of consolidation is controlled
by the coefficient of consolidation which, in turn, is governed by the hydraulic

conductivity and not by the electro-osmotic permeability. For a given
compressibility, the amount of consolidation depends on the magnitude of
negative pore-water pressure that can be developed at the end of consolidation. At
the end of the electrical treatment the effective stress developed during the
electro-osmotic consolidation relaxes gradually and some rebound occurs.
Under ideal conditions, and starting with no excess pore-water pressure, at the
end of the period of electro-osmotic treatment a maximum pore-water tension
occurs at the anodes and decreases linearly to zero at the cathode. Hence, the soil
properties vary between anode and cathode. In order to develop a more uniform
stress distribution and to increase the average shear strength of the soil, it may
prove useful to reverse the roles of the electrodes after a certain period of
treatment.
Two factors mitigate against the use of electro-osmosis on a large scale.
Firstly, overconsolidated clays tend to be fissured. If the ability of such soil to
transmit water along these fissures is similar to the electroosmotic permeability,
then it is unnecessary to use electro-osmotic treatment. Cheaper methods of
drainage, such as vertical drains, can be used instead. More seriously, electro-
osmosis only operates efficiently in saturated and near-saturated soils. It
therefore cannot be used to dewater a soil completely. However, Bjerrum et al.
(1967) noted that an excessive current density means that the soil around the
anodes dries out, thereby increasing the anode-soil resistance so that the process
becomes increasingly less effective. Cessation of the process, especially in clay
soils, permits reabsorption of water. This occurs at low hydraulic rates but,
nevertheless, after a certain period the soil reverts to its original state.
The mechanical properties of clay minerals vary—particularly those belonging
to the smectite family—according to the type of cations associated with them. If
a clay mineral contains a significant amount of weakly bonded cations such as
sodium or lithium, then it has a tendency to absorb and retain large quantities of
water. Accordingly, it possesses a low shear strength. By contrast, if the cations
have higher bonding strength—such as calcium or magnesium or, more
especially, iron or aluminium—the clay material absorbs less water and has a more
stable structure.
As cations associated with the clay minerals are exchangeable, the
introduction of solutions containing an excess of cations with higher bonding
strength gives rise to ion exchange, thereby enhancing the soil properties.
Introduction of stabilizing electrolytes into a clay soil is brought about by either
solution and dispersion of the anode material through electrolysis or by direct
introduction of additives at the anode.
Electrolysis normally involves decomposition of the electrolyte accompanied
by solution of anode material. Electrochemical stabilization due to anode
solution results from initial base exchange reactions on the surface of the clay
minerals, weakly bonded cations being replaced by strongly bonded cations, and
the formation of soil-cementing compounds produced by reaction between the
electrolyte and dissolved anode solution. Thus a lattice-like structure is built-up

around the clay particles. This greatly reduces the amount of settlement that
would result from the use of electro-osmosis on its own.
Addition of additives to the anode (in addition to the dissolved electrode
material) can increase the base exchange reaction, act as a catalyst in the
formation of cementitious material, and improve the quality and quantity of
cementitious material. To increase the base exchange reaction, most of the
additives used are organic and inorganic compounds of aluminium or calcium,
iron normally being present in the form of the electrode material. Of these the
most commonly used compound is calcium chloride. This intensifies dissolution
of the iron anode by the formation of ferrous chlorides, at the same time
producing calcium hydroxide which acts as a cementing agent.
An important disadvantage in the use of additives is the reduction in flow with
increase in electrolyte concentration. This may effectively limit stabilization
beyond a certain level.
6
Compaction and consolidation techniques
6.1
MECHANICAL COMPACTION OF SOIL
6.1.1
Introduction
Mechanical compaction is used to compact fills and embankments by laying and
rolling the soil in thin layers. In other words, it refers to the process by which
soil particles are packed more closely due to a reduction in the volume of the
void space, resulting from the momentary application of loads such as rolling,
tamping or vibration. It involves the expulsion of air from the voids without the
moisture content being changed significantly. Hence the degree of saturation is
increased. However, all the air cannot be expelled from the soil by compaction so
that complete saturation is not achievable. Nevertheless, compaction does lead to
a reduced tendency for changes in the moisture content of the soil to occur. The
method of compaction used depends upon the soil type, including its grading and
moisture content at the time of compaction; the total quantity of material, layer
thickness and rate at which it is to be compacted; and the geometry of the
proposed earthworks.
With a soil of a given moisture content, increasing compaction results in
closer packing of the soil particles and therefore in increased dry density. This
continues until the amount of air remaining in the soil is so reduced that further
compaction produces no significant change in volume. The soil is stiff and
therefore more difficult to compact when the moisture content is low. As the
moisture content increases it enhances the interparticle repulsive forces, thus
separating the particles causing the soil to soften and become more workable.
This gives rise to higher dry densities and lower air contents. As saturation is
approached, however, pore-water pressure effects counteract the effectiveness of
the compactive effort. Each soil, therefore, has an optimum moisture content at
which the soil has a maximum dry density (Figure 6.1). This optimum moisture
COMPACTION AND CONSOLIDATION TECHNIQUES 109
Figure 6.1 A compaction curve indicating maximum dry density and optimum moisture
content
content can be determined by the Proctor or standard compaction test (Anon.,

1976). Unfortunately, however, this relationship between maximum dry density
and optimum moisture content varies with compactive effort and so test results
can only indicate how easily a soil can be compacted. Field tests are needed to
assess the actual density achieved by compaction on site (Parsons, 1987).
Most specifications for the compaction of cohesive soils require that the dry
density that is achieved represents a certain percentage of some laboratory
standard. For example, the dry density required in the field can be specified in
terms of relative compaction or final air void percentage achieved. The ratio
between the maximum dry densities obtained in situ and those derived from the
standard compaction test is referred to as relative compaction. If the dry density
is given in terms of the final air percentage, a value of 5–10% usually is
stipulated, depending upon the maximum dry density determined from the
standard compaction test. In most cases of compaction of cohesive soils, 5%
variation in the value specified by either method is allowable, provided that the
average value attained is equal to or greater than that specified.
Specifications for the control of compaction of cohesionless soils either

require a stated relative density to be achieved or stipulate type of equipment,
thickness of layer and number of passes required. Field compaction trials should
be carried out in order to ascertain which method of compaction should be used,
including the type of equipment. Such tests should take account of the variability
in grading of the material used and its moisture content, the thickness of the
individual layers to be compacted and the number of passes per layer.
6.1.2
Types of compaction equipment
Sheepsfoot rollers are best suited to compacting cohesive soils in which the
moisture content should be approximately the same as the optimum moisture
content given by the modified AASHO compaction test. At least 24 passes are
required in order to achieve reasonably adequate compaction with a sheepsfoot
roller, and the layer of soil to be compacted should not be more than 50 mm
thicker than the length of the feet (Table 6.1). The resultant air voids content of
soil compacted by sheepsfoot rollers is rather greater than that of soil compacted
by smooth-wheel or rubber-tyred rollers.
Rubber-tyred rollers range in weight up to 100 tonnes. The large rollers are
usually towed by heavy crawler tractors. Heavy rubber-tyred rollers are not
recommended for initial rolling of heavy clay soils but are effective and
economical for a wide range of soils from clean sand to silty clay (Table 6.1). The
rubber-tyred roller produces a smooth compacted surface which consequently
does not provide significant bonding and blending between successive layers.
Also, the relatively smooth surface of a rubber tyre can neither aerate a wet soil
nor mix water into a dry soil. Rubber-tyred rollers are most suited to compacting
uniformly graded sands. When used to compact cohesive soils, rubber-tyred
rollers give the best performance when the soil is about 2–4% below the plastic
limit.
Maximum dry density can be obtained with rubber-tyred rollers by
compacting layers of soil, 125–300 mm in thickness, with from 4 to 8 passes.
These rollers generally can compact soil in less time and at lower cost than
sheepsfoot rollers.
Compaction by a rubber-tyred roller is sensitive to the moisture content of the
soil. For instance, if the moisture content is on the wet side of the laboratory
determined optimum, then this necessitates an increase in the number of passes of
a rubber-tyred roller in order to provide a given soil density. The optimum
moisture content required for compaction of a soil by a rubber-tyred roller occurs
at a higher degree of saturation than that for a sheepsfoot roller. This may be
detrimental in embankments where high construction pore-water pressures
cannot be tolerated. On the other hand, the construction of earth embankments in
regions where rainfall frequently occurs is expedited by the use of rubber-tyred
Table 6.1 Typical compaction characteristics for soils used in earthwork construction
(from BS6031:1981). The information in this table should be taken only as a general
guide. When the material performance cannot be predicted, it may be established by
earthwork trials. This table is applicable only to fill placed and compacted in layers. It is
not applicable to deep compaction of materials in situ.
rollers since they help seal the surface of the compacted soil and thereby reduce
infiltration.
Smooth-wheeled rollers are most suitable for compacting gravels and sands
(Table 6.1). In granular soils, the control of moisture content is important in that
it should be adjusted to the optimum moisture content. The depth of the layer to
be compacted is governed by the nature of the work as well as the weight of the
roller. Generally, however, individual layers vary in thickness, from 50 mm for
subgrades to 450 mm for the base of embankments. Smooth-drum rollers are not
recommended for compacting cohesive soils in earth dams because of their low
unit pressures and the smooth surface they produce.
Vibratory rollers may be equipped with rubber tyres, smooth-wheeled drums or
tamping feet (Lewis and Parsons, 1958). Vibration must provide sufficient force
(dead weight plus dynamic force) acting through the required distance
(amplitude) and give sufficient time for movement of grains (frequency) to take
place. The thickness of each compacted layer is governed by the vibration
frequency and weight of a vibratory roller. Lightweight, high-frequency rollers
obtain satisfactory densities in thin lifts, while heavyweight, low-frequency
rollers obtain satisfactory densities in thick lifts.
The speed and number of passes of a vibratory roller are critical, inasmuch as
they govern the number of dynamic load applications developed at each point of
the compacted fill. Increasing the number of passes increases the compactive
effort as well as the effective depth of compaction (Figure 6.2). Nonetheless,
after a few passes, a further increase in depth of compaction requires many
passes.
Vibratory rollers have been used successfully for compacting sand, gravel and
some cohesive soils (Toombes, 1969; Table 6.1). The frequency and deadweight
of vibratory rollers must be suited to the material being compacted. For instance,
heavyweight rollers with low-frequency vibrations are used to compact gravel;
light to medium-weight rollers with high-frequency vibrations are used for
sands; and heavyweight rollers with low-frequency vibrations are used for clays.
The best compaction is obtained with vibratory rollers when the soil is at or slightly
wetter than optimum moisture content.
Rammers can compact relatively thick layers of granular soils. However, when
they are used to compact clay, each layer should not exceed 225 mm in thickness.
The moisture content of soil compacted by a heavy rammer should be maintained
just below the optimum moisture content determined by the standard compaction
test. Because rammers have a low output their use is generally restricted to
special sites where rollers cannot be employed, such as trenches and behind
bridge abutments.
Figure 6.2 Relationship between dry density of 150 mm thick layer and number of passes
of vibrating rollers on a well-graded sand with a moisture content of 7.5% (after Lewis
and Parsons, 1958; Toombs, 1959) (with permission, the Transport Research Laboratory,
Berks)
6.1.3
Compaction and soil types
The engineering properties of soils used in fills, such as their shear strength,
consolidation characteristics and permeability, are influenced by the amount of
compaction they have undergone. Therefore, the desired amount of compaction
is established in relation to the engineering properties required for the fill to
perform its design function. A specification for compaction needs to indicate the
type of compactor, mass, speed of travel, and any other factors influencing
performance such as frequency of vibration, thickness of individual layers to be
compacted and number of passes of the compactor. Table 6.2 provides a choice
of compaction plant for the construction of earthworks. Procedures for the use of
compaction plant with different soil types are provided, as are the number of
passes and thickness of the layer designed to give a compactive effort capable of
achieving an adequate state of compaction with the more difficult soil conditions
likely to be encountered.
The compaction characteristics of clay are largely governed by its moisture
content (Hilf, 1975). For instance, a greater compactive effort is necessary as the
moisture content is lowered. It may be necessary to use thinner layers and more
passes by heavier compaction plant than required for granular materials. The
Table 6.2 Three methods specified for the compaction of earthwork materials (Anon.,
1986a)
* Rollers should be towed by track-laying tractors. Self-propelled rollers are unsuitable.

D=maximum depth at compacted layer (mm). N=minimum number of passes.
properties of cohesive fills also depend to a much greater extent on the

placement conditions than do those of a coarse-grained fill.
The shear strength of a given compacted cohesive soil depends on the density
and the moisture content at the time of shear. The pore-water pressures
developed while the soil is being subjected to shear are also of great importance
in determining the strength of such soils. For example, if a cohesive soil is
significantly drier than optimum moisture content, then it has a high strength
owing to high negative pore-water pressures developed as a consequence of
capillary action. The strength declines as the optimum moisture content is
approached and continues to decrease on the wet side of optimum. Increases in
pore-water pressures produced by volume changes coincident with shearing tend
to lower the strength of a compacted cohesive soil.
Furthermore, there is a rapid increase in pore-water pressures as the moisture
content approaches optimum. Nevertheless, compaction of cohesive soil with
moisture contents that are slightly less than optimum frequently gives rise to an
increase in strength, because the increase in the value of friction more than
compensates for the change in pore-water pressure.
The compressibility of a compacted cohesive soil also depends on its density
and moisture content at the time of loading. However, its placement moisture
content tends to affect compressibility more than does its dry density. If a soil is
compacted significantly dry of optimum, and is then saturated, extra settlement
occurs on loading. This does not occur when soils are compacted at optimum
moisture content or on the wet-side of optimum.
A sample of cohesive soil compacted on the dry-side of optimum moisture
content swells more, at the same confining pressure, when given access to
moisture than a sample compacted on the wet-side. This is because the former
type has a greater moisture deficiency and a lower degree of saturation than the
latter, as well as a more random arrangement of particles. Even at the same
degrees of saturation, a soil compacted on the dry-side tends to swell more than
one compacted wet of optimum. Conversely, a sample compacted wet of
optimum shrinks more on drying than a soil, at the same density, that has been
compacted dry of optimum.
When expansive clays must be used as compaction materials they should be
compacted as wet as is practicable, consistent with compressibility requirements.
Certainly it would appear that compaction wet of optimum moisture content gives
rise to lower amounts of swelling and swelling pressure. But this produces fill
with lower strength and higher compressibility. Therefore, the choice of moisture
content and dry density should not be based solely on lower swellability.
Consideration should be given to adding various salts to reduce the swelling
potential by increasing the ion concentration in the pore water.
A minimum permeability occurs in a compacted cohesive soil that is at or is
slightly above optimum moisture content, after which a slight increase in
permeability occurs. The noteworthy reduction in permeability that occurs with
an increase in moulding water on the dry-side of optimum is brought about by an
improvement in the orientation of soil particles, which probably increases the

tortuosity of flow, and by the decrease in the size of the largest flow channels.
Conversely, on the wetside of optimum moisture content, the permeability
increases slightly since the effects of decreasing dry density more than offset the
effects of improved particle orientation. Increasing the compactive effect lowers
the permeability because it increases both the orientation of particles and the
compacted density.
The moisture content has a great influence on both the strength and
compaction characteristics of silty soils. For example, an increase in moisture
content of 1 or 2%, together with the disturbance due to spreading and
compaction, can give rise to very considerable reductions in shear strength,
making the material impossible to compact.
Because granular soils are relatively permeable, even when compacted, they
are not affected significantly by their moisture content during the compaction
process. In other words, for a given compactive effort, the dry density obtained is
high when the soil is dry and high when the soil is saturated, with somewhat
lower densities occurring when the soil has intermediate amounts of water (Hilf,
1975). Moisture can be forced from the pores of granular soils by compaction
equipment and so a high standard of compaction can be obtained even if the
material initially has a high moisture content. Normally granular soils are easy to
compact. However, if granular soils are uniformly graded, then a high degree of
compaction near to the surface of the fill may prove difficult to obtain,
particularly when vibrating rollers are used. This problem usually is resolved
when the succeeding layer is compacted in that the loose surface of the lower
layer is also compacted. According to Anon. (1981 (a)) improved compaction of
uniformly graded granular material can be brought about by maintaining as high
a moisture content as possible by intensive watering and by making the final
passes at a higher speed using a non-vibratory smooth-wheeled roller or grid
roller. When compacted granular soils have a high load-bearing capacity and are
not compressible they are not usually susceptible to frost action unless they
contain a high proportion of fines. Unfortunately, however, if granular material
contains a significant amount of fines, then high pore-water pressures can
develop during compaction if the moisture content of the soil is high.
It is important to provide an adequate relative density in granular soil that may
become saturated and subjected to static or, more particularly, to dynamic shear
stresses. For example, a quick condition might develop in granular soils with a
relative density of less than 50% during ground accelerations of approximately 0.
1 g. On the other hand, if the relative density is greater than 75%, liquefaction is
unlikely to occur for most earthquake loadings. Consequently, it has been
suggested that in order to reduce the risk of liquefaction, granular soils should be
densified to a minimum relative density of 85% in the foundation area and to at
least 70% within the zone of influence of the foundation.
Whetton and Weaver (1991) demonstrated that intensive surface compaction
by a heavy vibratory roller (10 750 kg), operating at 30 Hz, can be used to
Figure 6.3 Applicability of deep compaction methods (after Mitchell, 1970, with
permission of ASCE)
compact granular fill, the required density to avoid liquefaction being achieved
by 10–12 passes at a forward speed of 0.3–0.6 m/s. The layer compacted was 1.
8m thick of saturated gravelly sand. The greatest density within the layer was
obtained within the depth interval 0.6–1.2 m, relative density exceeding 85%. The
lesser densification above was explained as due to insufficient overburden
pressure for the selected vibration level and compactive energy. This produces a
loosening effect due to larger vibration of particles which has been referred to as
overcompaction or overvibration. Whetton and Weaver maintained that the
maximum depth of improvement increases with roller energy but that the
location of the water table, fines content or presence of any hard material that
reflects vibrations can have a significant effect on the maximum depth and
effectiveness of densification (the former increases while the latter two decrease
the depth of densification). Nonetheless, the vibration energy from intensive
surface compaction attenuates fairly rapidly with depth and the degree of
densification drops accordingly. The compaction of layers of clean sand up to 3
m in thickness by heavy vibratory rollers had previously been described by D’
Appolonia et al. (1969) and Moorhouse and Baker (1969).
Thick deposits of soft soils or loosely packed cohesionless soils require
improvement in order to reduce the amount of settlement, which occurs on
loading, to acceptable amounts. Methods used to bring about in-situ compaction
of soils include precompression, compaction piles, vibrocompaction, dynamic
compaction and explosive compaction. Their applicability to various soil types is
shown in Figure 6.3.
6.2
PRECOMPRESSION
Precompression involves compressing the soil under an applied pressure prior to
placing a load (Aldrich, 1965). As such, it has proved an effective means of
enhancing the support afforded shallow foundations and commonly is used for
controlling the magnitude of post-construction settlement. Precompression is
well suited for use with soils which undergo large decreases in volume and
increases in strength under sustained static loads when there is insufficient time
for the required compression to occur. Those soils that are best suited to
improvement by precompression include compressible silts, saturated soft clays,
organic clays and peats.
Precompression is normally brought about by preloading, which involves the
placement and removal of a dead load (Johnson, 1970). This compresses the
foundation soils thereby inducing settlement prior to construction. If the load
intensity from the dead weight exceeds the pressure imposed by the final load,
then this is referred to as surcharging. In other words, the surcharge is the excess
load additional to the final load and is used to accelerate the process of
settlement. The ratio of the surcharge load to the final load is termed the
surcharge ratio. The surcharge load is removed after a certain amount of
settlement has taken place. The installation of vertical drains beneath the
precompression load helps shorten the time required to bring about primary
consolidation. Earthfill, and to a lesser extent rockfill, is the most frequently used
material for precompression.
As stated, the objective of preloading and the length of time it is applied is to
reduce the amount of settlement that occurs after the construction period has
ended (Figure 6.4). To this end the amount of preloading needed to give the
required settlement at final pressure in a certain time is ascertained or,
alternatively, the time required to produce a given amount of settlement under a
particular preload is determined. The rate of time of settlement for one-
dimensional primary consolidation can be obtained by using the method given by
Terzaghi (1943). However, there are certain situations in which the Terzaghi
theory does not apply, as, for example, where the strain profile is not constant
with depth (Mitchell, 1981). Where surcharging is involved in precompression
the degree of consolidation, Uq+∆q, deloped under final loading, q, and
surcharge, ∆q, can be derived from
(6.1)
where ∆H and ∆Hq+∆q are derived from
(6.2)
and
Figure 6.4 Principle of precompression using surcharge loading (i.e. load in excess of
final load to accelerate settlement). The length of time to achieve the required amount of
settlement has to be assessed, or alternatively the magnitude of the surcharge, ∆q, required
to ensure that settlement, ∆H, will be completed in time, t2, has to be determined. If the
surcharge is left in place until time t2, then it will give the same amount of settlement as
that which would occur under the final load at time t1.
(6.3)
respectively, and in which

H=thickness of the layer
e0=initial void ratio
Cc=compressibility index
σ'0=initial effective overburden pressure.
When the necessary degree of consolidation, U, has been determined for
assumed values of permanent loading, the corresponding time factor, Tv, can be
determined from Figure 6.5(a). The time, t, for surcharge removal is then derived
from:
(6.4)
where cv is the coefficient of consolidation. If rebound following surcharge

removal is to be kept to a minimum, then the amount of permanent load
preferably should not be less than approximately one-third that of the surcharge.
Secondary compression may cause significant amounts of settlement in some
soils (mainly organic clays and peat). Hence the effect of surcharge must also
Figure 6.5 (a) Time factor, Tv, versus degree of consolidation, U. Where the consolidating
layer drains from both the top and bottom, that is, the layer is open, then curve (1) is used.
Where drainage takes place in one direction only, that is, the consolidating layer is half-
closed, the curve used depends upon the initial excess pore-water pressure as shown in
(b). (b) Initial variations in excess pore water pressure in a half-closed layer. If the layer is
half-closed, then H/2 is used in eq. (6.4).
take account of secondary compression (Figure 6.6). The rate of secondary
compression appears to decrease with time in a logarithmic manner and its
amount is directly proportional to the thickness of a compressible layer at the
start of secondary compression. Johnson (1970) noted that secondary
compression appears to be due to shear stresses and that settlements therefore
beneath the edges of fills where the shear stresses are high may exceed values
beneath the central part of preload fills. An approximate procedure for reducing
the amount of secondary compression that occurs after construction involves
adding a secondary compression time settlement curve to the primary
compression curve in which the amount of secondary compression, ∆H2, is
derived from
(6.5)
where
Figure 6.6 Surcharge to reduce rate of secondary compression (after Bjerrum, 1972, with
permission of ASCE). It is assumed that o′0+q+∆q is well above pre-compression
pressure, pc, in each layer (where σ′0 is the initial effective stress) so that the
representative point C in the e-log p graph is on the virgin curve. Because secondary
compression may cause large amounts of total settlement in some soils (mostly organic
clays and peat), the effect of surcharge must be analysed with reference to this secondary
compression. If one considers the e-log p graph, the primary compression path is ABC and
secondary compression can be represented according to Bjerrum by CD. In most cases the
actual path is represented by the dotted line between B and D because secondary
compression starts during the primary compression period. A precompression project with
surcharge is represented by ABC′D′ (or ABD′) and surcharge removal is D′E. The degree
of secondary consolidation can be read on the set of e-log p curves giving the equilibrium
void ratio at different times of sustained loading (0.1 year, 1 year, 10 years, etc.). Hence,
due to surcharge, a much greater degree of secondary consolidation (100 years) can be
achieved than without: a state which is represented by point D (6 months)
Cs=the coefficient of secondary compression (it can be regarded as the ratio of

decrease in sample height to the initial sample height for one cycle of time on
logarithmic scale following completion of primary consolidation)
H=initial thickness of compressible stratum
ts=the useful life of the structure or time for which the settlement is significant
tc=time for the completion of primary consolidation.
In the majority of cases where precompression has been used, it has involved
soft fine-grained soils that were either normally consolidated or only slightly
overconsolidated. Precompression, as remarked, also has been used to treat peaty
soils (Sasaki, 1985). The time available for compression generally limits the
practice to relatively thin layers and to soil types that compress rapidly. Soil
undergoes considerably more compression during the first phase of loading than
during any subsequent reloading. Moreover, the amount of expansion or heave
following unloading is not significant.
The geological history of a site and details regarding the types of subsoil, its
stratification, strength and compressibility characteristics are of greater
importance as regards the successful application of precompression techniques
than they are for alternative methods. As can be inferred from above, of
particular importance is the determination of the amount and rate of
consolidation of the soil mass concerned. Hence sufficient samples must be
recovered to locate even thin layers of silt or sand in clay deposits in order to
determine their continuity. For example, the presence of thin layers of sand or
silt in compressible material may mean that rapid consolidation takes place.
Unfortunately, this may be accompanied by the development of abnormally high
pore-water pressures in these layers beyond the edge of the fill. This lowers their
shearing resistance ultimately to less than that of the surrounding weak soil.
Accordingly excess pore-water pressures may have to be relieved by vertical
drains.
If a preload fill extends over a long wide area, the influence of layers of sand
or silt may be of no consequence as far as accelerating the rate of consolidation
beneath the central area of the fill is concerned (Figure 6.7). There consolidation
of the soil may be attributable almost exclusively to drainage in a vertical
direction. In addition, high pore-water pressures may develop beyond the edges
of the fill. As a result, vertical drains or relief wells may be required beneath the
central part of large preload fills and beneath their edges.
The installation of drains beneath a preload accelerates the rate of settlement
in the soil. For example, Robinson and Eivemark (1985) described the use of
wick drains and preloading on soft clayey silt. The required settlements under the
preload fill were achieved one month after the construction of the preload.
Without wick drains a similar amount of settlement would have required up to
two years of preloading. Accelerated settlement brought about by sand drains
beneath a surcharged fill has also been reported by Moh and Woo (1987).
Precompression also can be brought about by vacuum preloading by pumping
from beneath an airtight impervious membrane placed over the ground surface
and sealed along its edges. In order to ensure the distribution of the low pressure,
a sand layer is placed on the ground beforehand. The ‘negative’ pressure, created
by the pumps, causes the water in the pores of the soil to move towards the
surface because of the hydraulic gradient set up. The degree of vacuum that can
be obtained depends on the pump capacity and airtightness of the seals (i.e.
plastic sheeting, the edges of which are buried in narrow trenches sealed with a
Figure 6.7 Excess pore-water pressure distribution with depth at times of degree of
consolidation U1 and U2. For a degree of consolidation Uq+∆q=U1, although ∆H1 is
achieved, there can still be values of excess pore-water pressure, ∆u1, much larger than ∆q,
in the central part of the layer. On the other hand, in the upper portion (and in the lower
portion when both boundaries are pervious) excess pore pressures are much less than ∆q.
Consequently, after removal of surcharge the upper and lower parts of the clay layer may
heave and the central part continues to undergo progressive settlement due to primary
consolidation. To eliminate further primary consolidation following the removal of
surcharge it is necessary to wait until the degree of consolidation is equal to U2. Surcharge
helps develop secondary consolidation settlement and the degree of secondary
consolidation is much higher than for simple precompression. Also, the rate of further
consolidation is much smaller. This procedure also assumes that excess pore-water
pressures are dissipated in due time and that effective stresses are able to induce proper
secondary consolidation
bentonite-cement mixture). Values of 60–70% vacuum are generally attained.
The consolidation process is similar to that obtained by preloading. It should,
however, be emphasized that the low pressure in the voids of the soil increases
the size of air bubbles which, in turn, may reduce the permeability of the soil. As
in preloading, the method can be improved by use of vertical drains (Woo et al.,
1989; Figure 6.8). It can also be used with surcharge loading, the vacuum first
being applied to enhance the behaviour of the soil with the surcharge being
placed subsequently (Tang and Gao, 1989).
The vacuum method is especially suited to very soft soils where a surcharge may
cause instability. Choa (1989) described the use of the vacuum method and band
Figure 6.8 Preloading by vacuum. In pressure diagrams σ′ is effective pressure and u is

pore-water pressure. The pressure diagram on the left is for ground conditions prior to
treatment; that on the right indicates the changes that occur when vacuum is applied,
notably the development of negative pore-water pressure and the increase in effective
pressure which lead to consolidation
drains to improve very soft to soft silty clay. A possible advantage of the vacuum
method is the use of cyclic preloading of the ground, which can lead to an
increased rate of settlement. The vacuum method can produce surcharge loads up
to 80 kN/m2.
Water (preponding) has also been used for preloading. Sometimes the ponds
may be lined. Water loading in storage tanks is often economical where tanks are
to be water tested and where the products that are subsequently to be stored
weigh less than water. For example, oil tank farms frequently are located on
marshy sites, and in such cases the total tank loads often impose stresses on the
soft foundation soils up to ten times their undrained shear strength. Accordingly,
controlled water tests have been used to preload foundation soils prior to
bringing the oil tanks into service.
Inundation of loess soils has been used in Russia in order to bring about collapse
of those types with a metastable soil structure (Evstatiev, 1988). This has
included localized shallow wetting of the ground over a number of months from
pipelines or shallow excavations where a constant water level is maintained on
the one hand and intense deep localized wetting on the other, brought about, for
example, by irrigation. If the flow rate in the latter case is high enough to
produce a continuous rise in the water table, then saturation of the collapsible
soil may occur within several months and the resulting settlement may be
extremely uneven. Compaction can be accelerated by the use of vertical sand
drains, especially if water is fed into them under pressure. This is the most cost-
effective method of treating great thicknesses of loess. However, it may cause
some post-treatment deformation which can delay construction, and fissures may
develop in the soil.
Consolidation is brought about by dewatering techniques (see Chapter 5). As

the water table is lowered this increases the loading on the soil by 9.81 kN/m2 for
every metre of lowering.
6.3
COMPACTION PILES
Soil can be compacted by driving piles into it. Compaction is achieved by simple
displacement. However, when the technique is employed in granular soils,
compaction is also brought about by the vibratory effect of driving piles, as well
as by displacement. Therefore the amount of compaction brought about by
compaction piles not only depends on their spacing and size but also upon the
effect of the vibration energy. This is influenced by the content of fines in a deposit
since transmission of vibrations is dissipated by the damping effect of fines. In
general, if the fines content exceeds 20%, the improvement will be reduced and
piles will need to be more closely spaced. Chung et al. (1987) suggested an
upper fines context of 35% if vibration was going to have an effect on
compaction. Compaction piles can be structural piles (timber or concrete) or sand
compaction piles.
While densification of loose sand can be brought about by driving any type of
displacement piles, sand piles are generally used when the sole purpose is
densification (Mitchell, 1970). Sand piles are installed by driving a hollow steel
mandrel with a false bottom to the required depth, filling the mandrel with sand,
applying air pressure to the top of the sand column, and withdrawing the
mandrel. This leaves a column of sand in the hole (Brons and De Kruijff, 1985).
During the installation of sand piles the pore-water pressures are increased.
However, due to the relatively close spacing of the sand columns and their high
permeability, the dissipation of excess pore-water pressures takes place at high
rates. The design of sand compaction piles is normally based on the stability of
the sand column, and assumes, firstly, that most of the load increase is carried by
the sand column which is supported by the surrounding soil, carrying a small
increase in load, and, secondly, that the piles and soil are subjected to equal
vertical deformation. The sand piles form columns on which the subsequent
structural load is concentrated. A well-compacted layer of soil is usually placed
on top of the sand piles to obtain a better load concentration on the piles by
arching. When densifying soils of low permeability below the water table, sand
piles also serve as drains.
The vibro-composer method is similar and involves driving a casing to the
required depth by means of a vibrator. A quantity of sand is then placed in the
pipe which is partially withdrawn as compressed air is blown down the casing
(Figure 6.9). Next the pipe is vibrated downwards to compact the sand pile. The
process is repeated until the ground surface is reached.
The formation of high-energy input, high-displacement compaction piles
involves boring a hole, 0.5 m in diameter, to the required depth in the ground by
Figure 6.9 Construction of compaction piles by the composer system

bailing from within the casing and then ramming sand from it with a drop-
hammer, delivering over 100 kJ of energy per blow (Figure 6.10). The tube is
withdrawn in short lifts and at each lift a charge of sand is rammed from the
tube. Not only is the sand pile, so formed, highly compacted but it has a
considerably greater diameter than that of the tube, thereby compacting the soil
around the pile.
The Franki technique of placing gravel compaction piles is similar and
includes a tube closed by a plug of gravel (or zero slump concrete) which is
driven into the ground by means of a drop-hammer striking the plug. During
driving, the soil is compacted radially around the tube. When the tube has been
sunk to the design depth an expanded base is formed by ramming a measured
quantity of gravel or dry concrete from the tube with a high-energy hammer. The
tube is then raised in lifts and further gravel is driven from the tube so forming a
continuous pile of compacted gravel. By measuring the quantity of gravel placed
in the pile, its diameter may be calculated and, by computing displacement, some
measure of the increase in the density of the surrounding soil may be
determined. Such piles are constructed on a predetermined spacing that has been
established in a test panel before construction begins.
Figure 6.10 Stages in the construction of displacement compaction piles. (1) Boring by
bailer with casing to full depth. (2) Pouring granular fill in increments of 2–3 m. (3)
Withdrawal of the casing partially and ramming the fill to the specified set; bottom of
casing should be at least 0.3 m below the top of the rammed fill. (4) Add further granular
fill and repeat stage (3) until the complete length of the column to ground level is formed
6.4
VIBROCOMPACTION
6.4.1
Vibroflotation in granular soils
Vibrations of appropriate form can reduce intergranular friction among loose
cohesionless soils and in this way bring about their densification (Table 6.3). The
action allows particles to be rearranged, unconstrained and unstressed, by
gravitational forces into the densest possible state. As the grains are rearranged
unstressed, no stress-strain readjustment occurs after compaction. Consequently,

compaction is permanent. The initial void ratios and compressibilities of
granular soils are reduced significantly as a result of vibroflotation and their
angles of shearing resistance are increased (Figure 6.11). This, in turn, enhances
their bearing capacity, reduces the amount of likely settlement due to loading and
increases the resistance of sandy soils to liquefaction. According to D’
Appolonia (1953), a relative density of 70% tends to be selected as the criterion
for adequate compaction. Compaction not only increases the relative density but
it enables the soil to carry heavy or vibrating loads with virtually no settlement
since the dynamic stresses involved are less than those produced by the vibroflot
(Figure 6.12). It also provides a significant reduction in the risk of settlements
during earth tremors.
Vibroflotation is best suited for densifying very loose sands below the water
table. The position of the water table is assumed to have no influence on the
densities produced. In fact maximum density can be obtained in clean free-
draining sands by vibration under either dry (water can be jetted from the
vibroflot) or saturated conditions. This is
Table 6.3 Effect of vibroflotation in cohesionless soils (after Mitchell, 1970)

Soil type Depth (m) Column spacing (m) Relative density (%)
Before After
Sand 7 2.0 43 80
Sand and gravel 4.5 – 63 85–95
Well-graded sand 6 2.4 47 79
Sand and gravel 4–5 2.1–2.4 7–58 70–100
Clean loose sand 4 2.3 33 78
Fine sand 5–6 2.0–2.4 0–40 75–93
Gravelly sand 6–9 – 0 80
Sandy gravel – – 33–80 85–95
Glacial sand and gravel 5 1.8–2.3 40–60 85–90
Well-graded till 3 3.0 50 75
Loose sand 7 1.9 Loose 80
Loose fine sand 6 1.8–2.3 Loose 80
because a quick condition is developed which reduces the shearing resistance of

the sand, thereby facilitating the movement of particles into a denser condition.
Liquefaction occurs within the immediate vicinity of the vibroflot, that is, up to a
distance of about 500 mm. Beyond this, liquefaction is incomplete because of
damping effects.
On the other hand, the presence of layers of clay, excessive amounts of fines,
cementation or organic matter in a soil can cause difficulties for vibroflotation. The
densities achieved and the radius of the zone of compaction decrease with
Figure 6.11 Effect of fines content on increase in penetration resistance by vibroflotation

(after Saito, 1977)
increasing silt, clay and organic content (Figure 6.13). Silt, clay and organic
material damp the vibrations created by the vibroflot, afford cohesion to particles
of sand and fill voids between particles. In this way the relative movement
between particles, necessary for densification, is restricted. Layers of clay also
reduce the radius of the zone of compaction. Furthermore, clay particles become
suspended in the wash water, forming a cake on the walls of the hole and mixing
with the backfill material.
The vibroflot, which is usually suspended from a crawler crane (Figure 6.14
(a)), can penetrate the ground under its own weight and energy but it is usually
assisted by jets of water or air emitted from its base (Figure 6.14(b)). The
primary function of jetting, however, is to provide support to the borehole during
treatment. Indeed, water is used whenever the borehole formed by the vibroflot
is potentially unstable or where groundwater flows into the hole. The water is
Figure 6.12 Relative density of clean sand at points midway between centres of vibration
as a function of probe spacing (after Thorburn, 1975)
allowed to overflow from the hole at the surface so that excess hydrostatic
pressure and outward seepage forces support the sides.
Compaction of granular soil occurs both during penetration and as the vibroflot
is extracted, when it is slowly surged up and down. During compaction,
downward movement of soil immediately surrounding the vibrator creates a
cone-shaped depression at the surface which has to be filled continuously with
selected granular fill material. The surface may be rolled at the end of treatment.
Several interrelated factors influence the densities that can be achieved by
vibroflotation (Brown, 1977). These include the type of equipment used, the
pattern and spacing of vibroflot centres (Figure 6.15), vibroflot withdrawal
procedures, the nature of the backfill and the quality of workmanship.
Furthermore, if sands and gravels are to be compacted, they should be
sufficiently permeable to permit the release of excess pore-water pressure as they
are subjected to vibration. In addition, granular sands should not contain more
than 20% fines, of which less than 3% should be active clay. The frequency of the
soil vibrations is the same as that of the vibroflot and in order to compact loosely
packed granular soils the amplitude and lateral reaction at the end of the
vibroflot must be maintained—the larger the amplitude, the larger the radius of
influence. The greater the power of vibroflot, the higher is the amplitude of the
vibrations. When the area compacted is loaded, the passive resistance of the
surrounding soil prevents lateral expansion.
Figure 6.13 Range of particle size distribution suitable for densification by vibroflotation
(after Brown, 1977, with permission of ASCE). (B—most suitable range)
The soil is compacted around the vibroflot to diameters usually between 2.4
and 3 m. By inserting the vibroflot in a pattern of overlapping compaction
centres, treatment of any required area can be carried out. A triangular grid
pattern of treatment centres is usually adopted, the spacing varying according to
the site conditions and required bearing capacity. As an example, spacing the
centres for vibroflotation 1.8 m apart may develop allowable bearing capacity of
340 kN/m2 compared with 180 kN/m2 for a pattern based on 2.7 m centres
(Figure 6.16). A square pattern with holes spaced at 1.8–2.25 m centres also may
be used for compaction. The fill material, which has to be added to compensate
for compaction, increases the dry unit weight of the soil by roughly 10–15%.
Most of the compaction takes place within 2–5 mins at any given depth of the
vibroflot in granular soils. It is uneconomic to try to obtain a higher degree of
densification by excessively increasing the time of vibration. In order to obtain
adequate compaction, it is necessary to supply enough backfill to transmit
vibrations from the vibroflot into the surrounding soil and to fill the void left as
the vibroflot is withdrawn. Accordingly, the rate of withdrawal of the vibroflot
must be consistent with the rate at which backfill material is supplied. If the
vibroflot is retracted in small lifts the maximum attainable densities are
achieved. For instance, if a vibroflot is raised in increments of 0.3 m and held for 30
Figure 6.14 (a) Vibroflot suspended from a crawler crane (courtesy of Cementation
Ground Engineering Ltd)
s this gives the same withdrawal rate as one raised in 1.2 m increments and held
for 2 mins. However, the 1.2 m lift gives more irregular compaction than the 0.3
m lift.
6.4.2
The terra-probe
Granular soils of low density can be compacted by driving and extracting a large
open-ended pipe, some 760 mm in diameter, in regular patterns in a deposit using
a vibratory pile driver. The pipe is 3–5 m longer than the desired penetration
depth. Densification of the soil takes place inside and outside the pipe. The
vibrations generated are vertical and the frequency of the vibrator can be varied.
Normally, however, the frequency is around 15 Hz. About 15 probes per hour
can be carried out at spacings of 1–3 m. The process does not require the
addition of backfill around the probe, as occurs in vibroflotation. The increase in
density of the soil means that it also undergoes settlement. Consequently, a
surcharge of soil usually is added to restore the soil to its original level. It is
advantageous to place the necessary surcharge before the compaction operation.
The final 0.9–1.5 m is compacted by surface rolling.
Figure 6.14 (b) Essential features of a vibroflot

6.4.3
Vibroreplacement
Clay soils are largely unaffected by induced vibrations. Nevertheless,
vibrocompaction is also used to treat clay soils. In this case stabilization is
brought about by columns of densely compacted coarse backfill (Figure 6.17).
These stone columns utilize the passive resistance of the soil, thereby increasing
its bearing capacity and reducing settlement (Figure 6.18). Stabilization is also
aided by the increased rate of dissipation of excess pore-water pressures due to
the presence of the columns and the gain in shear strength consequent upon radial
drainage (Figure 6.19). Radial drainage means that settlement occurs more
rapidly. Stone columns are also a means of reinforcing soil.
A large proportion of any load imposed at the surface is initially borne by the
stone columns, the remainder being carried by the soil (Greenwood, 1970;
Hughes and Withers, 1974). The vertical displacement of the tops of stone columns
within the range of working stresses is less than half the maximum radial strain
in the column. Only small radial strains are necessary to develop the minimum
passive resistance of the soil since the radial displacement of the soil associated
with the formation of the stone columns is considerable. The distribution of the
total load between the stone columns and the soil gradually changes until
equilibrium conditions are attained. The period of time over which this occurs is
Figure 6.15 Typical vibroflotation arrangement for footings (after Brown, 1977, with
permission of ASCE)
influenced by the ground and loading conditions and the spacing of the stone
columns.
Columns are constructed at numerous centres within clay. The columns are
typically 0.9 m in diameter but have varied from 0.5 to 1.5 m. The maximum
diameter depends on the properties of the clay. Each column can support a load
up to 300 kN.
In soft soils the vibroflot penetrates principally by water jetting in combination
with displacement which erodes a hole larger than the machine. The suspension
of the vibroflot from a crawler crane ensures that the hole produced is vertical.
The vibroflot penetrates the soil to the required depth and is then withdrawn.
About 1 m of granular fill is then placed in the bottom of the hole. This is
compacted, which tends to displace the surrounding soil laterally (Figure 6.17).
The procedure is completed in 1 m lifts up the hole so forming a column of very
dense hard granular material. The fill material should vary in particle size from
about 5 to 100 mm with not more than 15% material finer than 5 mm.
When backfill is placed with water still flowing up the annular space, vibration
transmitted through the backfill to the soil may cause local collapse of
Figure 6.16 Allowable bearing pressure as a function of probe spacing for footings
having widths varying from 1 to 3 m (after Thorburn, 1975)
surrounding clays, especially if they are sensitive. Collapsed material is removed
by the flowing water and the diameter of the hole grows as more backfill is
placed.
Considerable excess pore-water pressures are developed in saturated very soft
cohesive soils during the initial penetration of the vibroflot, and the soil moves
back into position behind the tip of the vibroflot during extraction. Furthermore,
very soft cohesive soil liquefies as a result of the high excess pore-water
pressures and the high stresses induced by the intense vibrations. It is therefore
necessary to form the hole by jetting water from the outlets at the base of the
vibroflot. This action flushes out the cohesive soil and prevents the development
of high excess pore-water pressures during initial penetration.
Vibroreplacement is commonly used in normally consolidated clays, soils
containing thin peat layers, saturated silts and alluvial or estuarine soils
(Figure 6.20). Stone columns have been formed successfully in soils with
undrained cohesive strength as low as 7 kN/m2, but normally it is not used in
Figure 6.17 The vibrocompaction process (after Baumann and Bauer, 1974, Canadian
Geotechnical Journal Vol 11.)
Figure 6.18 The effect of stone columns on anticipated foundation settlement (after
Greenwood, 1970)
soils with undrained shear strengths of less than 14 kN/m2 because of the low radial
support afforded the stone columns. Vibro-replacement is not recommended as a
method of treatment for deep deposits of highly organic silts and clays or thick
deposits of peat. Thorburn (1975) maintained that vibroreplacement cannot be
relied upon to strengthen made ground which contains large pockets of
household refuse or masses of organic matter because the position and extent of
Figure 6.19 Loading test results on treated and untreated sand

these materials cannot readily be established. His experience had shown that
radial strains undergone by stone columns due to the presence of these highly
compressible materials can cause significant non-uniform vertical foundation
displacements. Made ground containing heavy demolition spoil or timbers can
prove difficult or impossible to treat effectively by vibroreplacement.
6.4.4
Concrete columns
Concrete columns, normally 450 mm in diameter and compacted by a vibrator,
can be used to support structures on weak cohesive, organic or loose granular
soils and fills as an alternative to vibroreplacement. The technique involves the use
of a vibrator which contains a special hollow shoe that houses a tremie pipe
Figure 6.20 Treatment of mixed clays and sands: (a) stone columns stiffening clay lenses;
(b) compacted sand raft over soft clay
down which the concrete is pumped. The pipe is connected by hose to a trailer-
mounted concrete pump (Figure 6.21). The tremie pipe is initially charged with
concrete which is kept under pressure so that the shoe does not become plugged
with soil as the ground is penetrated. Vibrating at 3000 Hz, the vibrator is slowly
driven to the required depth. In the process weak cohesive soils are improved by
lateral displacement and loose sandy soils are compacted. Once at the required
depth the vibrator is withdrawn about a metre while pressurized concrete is
tremied in at the toe. The vibrator, still discharging concrete, is then driven back
to the original depth to create a bulbous bearing end some 600 mm in diameter.
The vibrator, still emitting concrete, is withdrawn at a slower speed than the
pumping flow so as to maintain pressure in the concrete column. This forces
concrete into the surrounding soil, thus forming a column with a diameter larger
Figure 6.21 Construction sequence of vibrated concrete column
than the vibrator. Once grade level is reached the concrete is shut off and the
vibrator is briefly re-immersed in the concrete to finish off the column.
6.4.5
Vibrodisplacement
A dry technique of forming stone columns in clays is also widely used. The dry
method is employed only where the hole remains stable and where there is no
risk of groundwater running into the hole. It is most effective in partially
saturated firm clays. The vibroflot penetrates by shearing and displacing the soil
around it. Because the vibroflot fits tightly into the borehole there is no annular
space into which backfill may be tipped. Consequently the vibroflot has to be
removed so that backfill can be placed into the hole. In saturated soft soils the
vacuum created as this happens may lead to caving of the sides of the hole.
Compressed air is circulated from the tip of the vibroflot to try to prevent this
occurring. However, compressed air should not be used if there is standing water
in the hole. The combination produces a clay slurry which mitigates against the
formation of a sound column.
The shearing action that occurs as soil is displaced as the vibroflot enters the
ground in the dry technique tends to break down the structure of the soil around
the hole. This may result in the gravel particles in the backfill not being in
mutual contact, which is essential for column stability. Shearing displacement
also may lead to any free draining layers that are intersected by the column
becoming blocked.
Vibrodisplacement accordingly is restricted to strengthening of insensitive
soils with an undrained strength of at least 20 kN/m2 and which, as stated, are
unlikely to collapse into the hole.
Displacement is responsible for an increase in the strength of the clay between
columns. This does not happen during vibroreplacement. The increase can be up
to 1.5 times the original strength for typical column spacing of 1.5 m.
A novel method of using vibrodisplacement to treat peaty soils has been
described by Bevan and Johnson (1989). Obviously, conventional
vibrodisplacement is not used to treat soils that include significant thicknesses of
peat. This is because construction of good continuous columns is difficult
through peat but even if these can be formed satisfactorily the surrounding peat
offers little lateral support. When load is applied, excessive bulging of the
column occurs into the peat. Large and unpredictable settlements then follow. In
order to avoid this occurring, trenches are excavated with a minimum of 1 m width
and extending 0.5 m below the base of the peat. The trench is then lined with a
geogrid which is fixed at the surface. Next granular fill is placed in the trench,
the lower 250 mm being 50 mm single size, the rest being filled with sand and
gravel. Lastly, the fill is compacted by vibrodisplacement. In this way light or
moderately loaded structures may be supported on conventional reinforced strip
footings with reinforced ground-bearing floor slabs.
6.5
DYNAMIC COMPACTION
Dynamic compaction is carried out by repeatedly impacting the ground surface
by dropping a pounder from a given height from a heavy duty crane at a rate of
one blow every 1–3 mins (Figure 6.22). Usually the blows are concentrated at
specific locations, the distances between the centres of impact frequently ranging
between 4 and 20 m, set out on a grid pattern. The energy per blow is chosen to
maximize penetration of the resultant stress impulses (Figure 6.23). Several
passes of tamping are required to achieve the desired result. Substantial
compaction results, thereby reducing the total and differential settlement that
may occur after the erection of structures and permitting the use of spread
footings (Gambin, 1987). Dynamic compaction has been used to densify a wide
range of soils from organic and silty clay to loosely packed coarse-grained soils
and fills (Menard and Broise, 1975). The energy required to achieve a given
result increases with the amount of fines in the ratio 1:3 when going from coarse
gravel to silty clay.
Dynamic compaction started with weights of about 8 t and drop heights of a
maximum of 10 m. Since then the general trend has been to utilize ever-
increasing weights and drop heights. At present a machine is in use which has a
capacity of lifting 200 t and has a drop height of 40 m (Figure 6.24). It can
Figure 6.22 Treatment of suspect ground at Ty Coch, Wales, by dynamic compaction

(courtesy of Menard Techniques Ltd)
compact a thickness of 50–60 m of clayey silt. However, dynamic compaction
does generate shock waves. Normally the vibration frequency is around 5 Hz,
and usually lower than 12 Hz, which represents no danger to existing buildings.
Nonetheless, large pounders appear to increase the amplitude of vibrations
significantly. The threshold vibration velocity above which damage may occur in
buildings is normally taken at 50 mm/s. Even so, damage has occurred at values
below this.
An extensive site investigation must be carried out prior to the commencement
of dynamic compaction in order that the energy requirements and number of
tampings to bring about the necessary improvement in ground conditions can be
assessed. A site that is to be treated by dynamic compaction may have to be
subdivided into different zones which have different requirements as far as drop
height, weight of the tamper and compaction energy are concerned. This depends
on the depth variation of loose soil and on varying requirements placed on the
site with regard to settlement. The depth of influence of dynamic compaction
depends on impact energy, the damping properties of the soil, the shape of the
tamper, the presence of gas in the pores of the soil and dissipation of pore-water
pressure.
Figure 6.23 Depth of influence as a function of impact energy for dynamic compaction
(after Mitchell, 1981)
Dynamic compaction is carried out in several phases (Figure 6.25). The
spacing between the points of impact is large in the first pass and is reduced
successively in subsequent ones (Hansbo, 1978). The shock waves and high
stresses induced by dropping the pounder result in the voids in the soil being
compressed, together with partial liquefaction of the soil and the creation of
preferential drainage paths through which pore water can be dissipated. This is
especially the case in fine sands and silts. The reduction in the void ratio brings
about an increase in soil strength and bearing capacity (Figure 6.26). The
modulus of deformation of the soil (assessed by the Menard pressuremeter) may
be increased by two to ten times which, in turn, means that settlement due to a
structural load is reduced.
It is essential to compact the soil at depth first, the soil at the surface being
compacted at the end of the operation. It also is important that the depth of the
imprint does not exceed half the width of the pounding mass, otherwise an
appreciable portion of energy is wasted in kneading the soil and is lost for
compaction; hence, it is necessary to have three or four types of tamping weights
available on site.
Figure 6.24 Menard Giga Machine drops 200 tonnes from 22 m in free fall and can
compact to a depth of 40 m, Nice airport, 12 m of sand fill resting on alluvium (courtesy
of Menard Techniques Ltd)
After each pass the craters formed in the ground by tamping are filled with
soil (ideally well-graded granular) before the next pass. At the end of the
compaction procedure the ground is given a final ‘ironing’ pass in which the
weight is dropped from 1 to 2 m at spacings smaller than the width of the weight.
In practice saturated or partially saturated soils and soft peaty or organic
deposits are not subjected to direct dynamic compaction. The standard procedure
involves placing a metre or so of fill, such as a sand blanket, over the soil to be
treated prior to its compaction. The stress wave produced by the pounder is then
somewhat attenuated by the time it reaches the saturated soil. This blanket serves
as a bearing layer for the crane and accelerates drainage of any pore water
expelled to the surface due to excess pressures. The tamping process both places
and mixes the fill into the soft soil beneath, the fill being made up between
tamping passes.
In the case of saturated soils, silt seems to form the lower safe boundary for
compatible materials with regard to grain size. When saturated silty soils are
subjected to dynamic compaction, the energy per pass must be limited and
several days must be allowed between passes. This permits the dissipation of
excess pore-water pressure which develops as the density of the soil is increased.
In the case of clayey soil it is recommended that a temporary surcharge should be
Figure 6.25 Ground response with time after successive coverages of dynamic
compaction (after Menard and Broise, 1975). Time between passes varied between one
and four weeks, according to soil type. (1) Applied energy in t m/m2; (2) volume variation
with time; (3) ratio of pore-water pressure to initial effective stress; (4) variation of
bearing capacity
placed on the soil before it is compacted. If dynamic compaction is carried out in
soft clays in association with vertical drains (such as the Geodrain, which can
withstand severe dynamic treatment), then this facilitates compaction.
A new, smaller machine has been introduced for ground compaction which
consists essentially of a large hydraulic pile hammer mounted on a crawler crane
(Figure 6.27). Although it can only compact the ground to a depth of some 4 m
compared with conventional dynamic compaction methods, which can reach
around 20 m, it can be used on small sites near existing buildings, which
dynamic compaction cannot. The soil is compacted by a 7 t hammer bearing down
on an anvil which is connected by an articulated joint to a foot at its base. The
Figure 6.26 Results of pressuremeter tests indicating improvement in strength of ground

treated by dynamic compaction
hammer delivers 30 blows per minute. This system ensures that a high
percentage of the energy is transferred to the soil. It can be used to improve sites
where light structures are to be erected.
6.6
EXPLOSIVE COMPACTION
Explosive compaction tends to be restricted to loosely packed granular soils
which contain less than 20% of silt or 5% of clay. For instance, the relative
densities of loose sand can be improved by 15–30%. Clay pockets in sands
reduce the efficiency of blasting operations. Sudden shocks or vibrations in
saturated loose granular soils, cause localized spontaneous liquefaction and
rearrangement of the component grains. The load is temporarily transferred to
the pore water and the grains are repositioned, adopting a much denser state of
packing. This reaction is aided by the weight of the soil above. A saturated
condition is very desirable for more uniform propagation of shock waves
(Carpentier et al., 1985), otherwise compaction by blasting may be somewhat
Figure 6.27 Tamping rig attached to crawler crane

erratic. Indeed, the pore water acts as a lubricant facilitating the rearrangement of
grains. The reduction of the void space in soil results in the displacement of a
large volume of water, the quantity depending on the depth, original void ratio
and amount of densification attained.
As can be inferred from the above, the most effective results of explosive
compaction are obtained in soils that are either dry or completely saturated. As
the water content of a soil is reduced from 100% saturation to the absorbed water
content around each grain of soil, the increase in capillary tension between the
grains means that only a lower state of density can be achieved. This also
becomes more pronounced as the soil becomes finer. The effect of capillary
tension can be overcome by ponding or flooding the area in question. The
presence of only a small amount of gas in the soil can give rise to significant
damping of the shock wave.
The distribution of charges, the delay in detonation between them, and their
size, should be based upon the required density of the material and the results of
field tests. Compaction appears to extend deeper in a layer of soil when small
delays are used between charges, that is, only a few seconds. But the amount of
compaction is less than with larger intervals. The development of craters at the
surface of the soil must be avoided. In addition to the desired amount of
densification, the size of the charge is influenced by the type of soil, its thickness,
the position of the water table, the dissipation of pore-water pressure, the
nearness of structures or open bodies of water, the spacing and overlapping
Figure 6.28 (a) CPT tests D62 and D63 in zone B before blasting; (b) minimum and
maximum qc values of tests S10 to S13 after blasting with tests D62 and D63 before
blasting (after Carpentier et al., 1985)
effect of charges, and type of explosives used (Pruch, 1963). If compaction is

brought about by a series of explosions, then the amount of compaction

following each round of blasting is usually less than that of the one preceding.
The maximum depth for effective treatment is limited by the difficulty of
placing concentrated charges of sufficient magnitude to generate a shock wave
large enough to bring about liquefaction in the soil. Furthermore, as depth
increases, so do the effective stresses and strength. Hence the required disruptive
stress is greater and the radius of influence decreases. Experience indicates that
an explosive charge should be placed at a point approximately two-thirds below
the centre of the mass to be densified. It may be necessary to place separate
charges in the same blast-hole when granular soils are separated by cohesive
soil. In fact thin lenses of clay in granular soils can adversely affect the degree of
compaction achieved, reducing it to almost nil (Figure 6.28).
Spacing of blast-holes ranges from 3 to 7.5 m and is governed by the depth of
the deposit, the size of the charge and the overlapping effect of adjacent charges.
In saturated soils, spacing closer than 3 m should be avoided (Figure 6.29) unless
carefully investigated for safety. The charges are placed in holes that are formed
by jetting with the aid of casing, and are 50 mm or so in diameter. Then the holes
Figure 6.29 Arrangement for blasting. The charges i, i+1, i+2 being detonated in the
second phase (after Carpentier et at., 1985)
are backfilled. The presence of a surface crust of impervious material inhibits the
expulsion of excess pore water from the soil. Hence, it must be removed,
disturbed or fractured.
Immediately after the blast the ground surface lifts and is fractured. Gas and
water escape from openings, similar to sand boils, which appear at the surface.
This may occur over periods ranging from minutes to hours. Primary settlement
of the surface follows the initial upheaval more or less immediately and
settlement proceeds for some time as the grains of soil continue to be rearranged
by the excess pore water that is being expelled. Little densification results in the
upper metre or so of soil. This zone, therefore, is compacted by a heavy large
mechanical vibrator, where possible.
Blasting has been used to bring about deep compaction in loess soil. Holes are
sunk to the required depth and then charged with explosives, the uppermost
cartridges being only 500–700 mm below the top of each hole. The charges are
detonated separately in each hole, with intervals between blasts of at least 1 min.
The use of preliminary flooding together with deep explosions also has been
used to compact loess soil (Litvinov, 1973). Ditches 0.2–0.4 m wide by 4–6 m
deep are excavated around the site to isolate the upper zones of the soil in the
Figure 6.30 Loess compaction by hydro-blasting
area to be flooded. Flooding takes several days and may be accomplished by the
aid of drains, a few metres apart, sunk into the soil (Figure 6.30). After wetting
the soil to a water content exceeding its liquid limit, explosions are set off in
boreholes spaced every 3–6 m within the soil. This accelerates compaction since
the wet soil is violently shaken which causes the metastable structure to collapse,
and settlement of the ground surface then follows. The rate of compaction is some
12 times faster than can be achieved simply by flooding loess, and the degree of
compaction obtained is from three to four times greater. Furthermore, the soil is
evenly compacted. Reductions in the porosity of soil can be from 50 to 33%.
7
Soil reinforcement and soil anchors
7.1
REINFORCED EARTH
7.1.1
Introduction
Reinforced earth is a composite material consisting of soil containing reinforcing
elements which generally comprise strips of galvanized steel or plastic geogrids
(Figure 7.1). Soil, especially granular soil, is weak in tension but if strips of
material providing reinforcement are placed within it, the tensile forces can be
transmitted from the soil to the strips. The composite material then possesses
tensile strength in the direction in which the reinforcement runs. The
effectiveness of the reinforcement is governed by its tensile strength and the
bond it develops with the surrounding soil. Both the shear strength of the soil and
the bond strength with the reinforcing elements are frictional and are thus
directly related to the normal effective stress distribution. Hence the presence of
reinforcement noticeably improves the mechanical properties of granular soil
and, depending on the confining pressure, σ3, two modes of failure can be
observed (Figure 7.2). The effectiveness of this composite material depends upon
the size, geometry and type of loading of the structure, as well as the types of
materials and drainage (Schlosser, 1987).
Reinforcing elements can be made from any material possessing the necessary
tensile strength and be of any size and shape which affords the necessary friction
surface to prevent slippage and failure by pulling out (Figure 7.3). They should
also resist corrosion or deterioration due to other factors. The corrosion/
durability of soil-reinforced structures has been discussed in detail by Elias
(1990). Steel or aluminium alloy strips, wire mesh, steel cables, glass-fibre-
reinforced plastic or polymeric geosynthetic materials (see Chapter 8) have been
used as reinforcing elements.
Figure 7.1 Reinforced earth system

Grid reinforcement systems consist of polymer or metallic elements arranged
in rectangular grids, metallic bar-mats, and wire mesh. The two-dimensional grid-
soil interaction involves both friction along longitudinal members and passive
bearing resistance developed on the cross-members. Hence grids are more
resistant to pull-out than strips, however, full passive resistance is only
developed with large displacements of the order of 50–100 mm (Schlosser,
1990). The most frequently used polymeric materials in earth-retaining structures
are high-density polyethylene and polypropylene grids (see Chapter 8). Several
systems of bar and mesh reinforcement have been developed. For example, crude
grids have been formed by cross-linking steel-reinforcing bars to form a coarse
welded-wire bar-mat. Welded-wire reinforcing mesh similar to that used for
reinforcing concrete has also been used to reinforce granular soil, especially in
the United States.
Attempts have been made to reinforce soil by including fibre within it. The
fibres have included geotextile threads, metallic threads and natural fibres such
as bamboo (Mitchell and Villet, 1987). A recent innovation is a three-
dimensional reinforcement technique which consists of mixing granular soil
backfill with a continuous polymer filament with a diameter of 0.1 mm and a
tensile strength of 10 kN (Leflaive, 1988; Schlosser, 1990). Approximately 0.1–
0.2% of the composite material consists of filament, resulting in a total length of
reinforcement of 200m per cubic metre of reinforced soil. The technique has
been used in soil-retaining walls and has demonstrated both high bearing
capacity and resistance to erosion. However, there are difficulties associated with
the efficient mixing of fibres with soil. In addition, randomly distributed
SOIL REINFORCEMENT AND SOIL ANCHORS 153
Figure 7.2 Triaxial test on reinforced sand (after Schlosser, 1987). At low values of
confining pressure, failure occurs by slippage of the reinforcements. The failure curve can
be approximated by a straight line passing through the origin. The strength of the
reinforced sand can be represented by an apparent friction angle, a, which is greater than
, the internal friction angle of the sand. The apparent friction angle is directly dependent
on the density of the reinforcement. The failure curve is a straight line parallel to the
failure line of the sand. The strength of the reinforced sand then can be represented by an
internal angle of friction, , and cohesion, c. The value of cohesion is directly
proportional to the density of the reinforcement and to it tensile strength
polymeric mesh elements have been included in soils, the meshes interlocking
with the soil particles to strengthen the soil. The ductility and permeability of the
soil are not reduced and a relatively homogeneous composite is produced. The
ribs of individual mesh elements interlock with groups of soil particles to form
an aggregation of partiicles, and adjacent aggregations interlock to form a
coherent matrix (Figure 7.4). Numerous types of mesh elements are now being
tested in a range of soil types and the practical problems of mixing them with
different soils in various situations are being investigated (McGown et al.,
1985).
Generally a limited range of soil types have been used as a fill in reinforced earth
structures. Indeed, the first specifications referred to clean granular materials.
For example, Vidal (1969) maintained that ideally the soil should be composed
mainly of granular material with a certain amount of clay. The amount of clay, it
was suggested, would vary with the type of reinforcement used, the general
condition being that sufficient friction should exist between soil and
reinforcement to generate the necessary tensile stresses in the reinforcement.
However, the type of fill used is also influenced by the requirements of the
Figure 7.3 Pull-out tests of high adherence and smooth strips (after Schlosser, 1987). The
peak and residual values of the apparent coefficient of friction depend on the density of
the backfill, the nature of the surface of the reinforcement and the overburden pressure. In
dense granular soils, the values of the apparent coefficient of friction, f*, are usually
significantly greater than the values obtained from direct shear tests. This is mainly due to
the fact that, under the effect of the shear the soil tends to increase in volume and the
denser the fill the more dilatant it will be for a given stress. This volume change is
restrained by the surrounding soil and the confining effect results in an increase of the
normal stress exerted on the reinforcement and consequently in a high value of the
apparent coefficient of friction. The dilatancy effect in a soil can be enhanced by forcing
more material to be sheared during the pull-out of the strip. In the case of high adherence
(HA) strips this is achieved by the ribs. Another interesting effect of shearing a larger
quantity of soil is that the peak, which occurs for a slightly larger displacement, is much
flatter. This indicates that the value of the coefficient of friction used in the design of
structures is the maximum value for high adherence reinforcements and the residual value
for smooth reinforcements. For a given density, a soil will be less dilitant if the confining
pressure is higher. This explains why the favourable effect of dilatancy on the apparent
coefficient of friction decreases when the average overburden stress ( H) increases. In
the case of high adherence reinforcements, the ribs force the shear surface to be within the
soil. Thus, under high overburden stress, the coefficient of friction approaches tan ,
being the internal friction angle of the soil. In the case of smooth strips, under high
overburden stress, the coefficient of friction approaches tan ψ, ψ being the soil
reinforcement friction angle. The value of tan ψ is approximately equal to 0.5 tan . Soil
having a soil reinforcement friction angle of less than 22° should be rejected
reinforced earth structure. As Jones (1985) pointed out, the type of fill used for
constructing reinforced earth retaining walls needs to be of better quality than
that used for a reinforced earth embankment.
Among the advantages of granular fill are the fact that it is free draining and
non-frost susceptible, as well as being virtually non-corrosive as far as the
reinforcing elements are concerned. It is also relatively stable, in that its use in
reinforced earth more or less elimates post-construction movements.
It is also necessary to provide some form of barrier to contain the soil at the
edge of a reinforced earth structure (Figure 7.5). This facing can be either
Figure 7.4 The interlock mechanism for mesh elements (after McGown et al., 1985): (a)
interlock with groups of particles; (b) interlock of adjacent aggregations
flexible or stiff but it must be strong enough to retain the soil and to allow the
reinforcement to be fixed to it.
Reinforced earth is now widely employed, primarily because the use of
prefabricated elements speeds up the construction process and because of the
lower cost of the technique compared with traditional methods. For example,
reinforced earth walls cost about half as much as cantilever or crib walls. Yet
another reason is the mechanical stability of the material.
Because reinforced earth is flexible and the structural components are built at
the same time as backfill is placed, it is particularly suited for use over
compressible foundations where differential settlements may occur during or
soon after construction. In addition, as a reinforced earth wall uses small
Figure 7.5 (a) Diagram showing ‘concertina’ type facings which were used particularly in
earlier reinforced earth. (b) Concrete panels used on a reinforced earth retaining structure
on the right-hand abutment of Clyde Dam, New Zealand
Figure 7.6 (a) Reinforced earth dam; (b) Stabilization of a slope with reinforced earth
prefabricated components and requires no formwork, it can be readily adapted to

required variations in height or shape.
Reinforced earth has been used in the construction of embankments, retaining
walls and bridge abutments for transportation applications; for industrial
structures including material processing and storage facilities, and foundation
slabs; for hydraulic structures such as sea walls, flood-protection structures,

sedimentation basins and dams (Figure 7.6); and for stabilization and
reinstatement of failed slopes (Figure 7.6; Mitchell and Villet, 1987; Mitchell
and Christopher, 1990; O’Rourke and Jones, 1990).
7.1.2
Load transfer mechanism and strength development
The load transfer mechanism between the soil and reinforcing element is
governed by the limiting friction that is developed at their interface. Because the
reinforcing action is derived from friction, it must be determined whether this
friction exists without sliding between grains and reinforcement. Andrawes et al.
(1980) indicated that sliding takes place if there is a tendency for the stress
obliquity to increase above that corresponding to the limiting friction-adhesion.
Reinforced earth behaves in an elastic manner that is capable of withstanding
both internal and external forces.
Because of friction, the upper and lower parts of a layer of soil enclosed
between adjacent strips of reinforcement are held by the reinforcement.
Consequently, if the strips are placed sufficiently close together, then the
complete layer of soil will be laterally restrained and the maximum strain
experienced by the soil in the direction of the reinforcements will approximate to
the strain in the reinforcing strips. Normally the value of Young’s modulus of the
reinforcement is much higher than that of the soil, which means that the resulting
strains in the soil are so small that it is essentially at rest.
Reinforced earth containing granular fill can be made to resist almost any
pressure since the grains usually have a high crushing strength and the
reinforcing strips can be placed inside the mass in order to resist external
conditions. Furthermore, deformations in the foundation will not give rise to
fissures or breaks in such reinforced earth. In such cases the particles slide along
the reinforcing strips and thereby achieve another state of equilibrium, generating
another stress distribution in the strips. As long as the reinforcing elements can
withstand these new loads, the whole structure remains stable. Schlosser and
Guilloux (1981) mentioned that many reinforced earth retaining walls built on
slopes had been subjected to large deformations and had remained intact. The
behaviour of such walls has demonstrated the great flexibility and good
performance of reinforced earth materials. Indeed, because reinforced earth can
suffer large deformations without rupture it is able to resist external forces such
as earthquakes.
Relative extension of the soil compared with the reinforcing strips becomes
less for higher walls and their corresponding higher stresses. Hence for higher
structures the effective lateral stress is reduced and approaches an active state.
Theoretically there is no height limit to reinforced earth structures.
In the case of reinforced earth walls, the wall facing and reinforced earth are
considered as a coherent block structure in stability analyses with active earth
Figure 7.7 Active and resistant zones in a reinforced earth wall
pressure acting behind the block. The stability of the block is then analysed
against sliding, overturning, bearing capacity failure and deep-sealed failure
(Manfakh, 1989). The locus of minimum tensile forces in the reinforcements
define active and resistant zones within reinforced earth. The effective length of
the reinforcements, le, is the length to the right of the active core in Figure 7.7. A
factor of safety of at least 1.5 is usually required against strip pullout. The
minimum tensile stress in the reinforcement must be less than that allowable for
the material.
Internal failure can occur only as a result of loss of friction between soil and
reinforcement, or by tensile failure of the reinforcing strips. The shear strength of
the fill is fully mobilized if the reinforcement breaks. Over stressing could lead
to tensile failure of the reinforcement. Determination of sliding shear resistance
between soil and reinforcement is not absolutely critical since slippage gives rise
to a redistribution of stress and slow deformation of the mass.
In order to obtain the most benefit from reinforced earth, reinforcing elements
should be laid in the direction of the principal strains within the soils. In most
reinforced earth structures that have been constructed, the reinforcing elements
are laid parallel to each other. Although this has proved a reasonable construction
expedient, it means that not all inclusions are aligned in the directions of
principal tensile strains or occur in zones of maximum tensile strains for the soil
under the same stress condition Thus the effectiveness of the individual layers
must vary as their orientations and locations deviate from their optimums. In fact,
weakening may occur where the orientation of a layer approaches that of the zero
extension lines of the unreinforced soil and the friction between an inclusion and
soil is less than that for the soil alone.
7.1.3
Soil types and reinforced earth
Two main factors influence the value of soil-reinforcement friction, namely, the
internal angle of friction of the fill material and the roughness of the surfaces of
the reinforcing elements. As remarked above, one of the most significant items is
the relative volume of the fine-grained fraction of the fill to that of the granular part,
since, as the former portion increases, so the internal friction angle decreases.
However, as long as the fine-grained fraction is small, the number of grain
contacts in the granular skeleton does not vary and so the value of the internal
angle of friction remains more or less the same. Schlosser (1987) suggested that,
ideally, fill material should not contain more than 15% (by weight) of material
smaller than 0.015 mm. He further suggested that the internal friction angle of
the saturated and consolidated fill material must be greater than 25°. The
assessment of internal friction is not required if no more than 15% (by weight) of
the fill material is smaller than 0.08 mm.
Special attention to moisture-density relationships is required when the
percentage finer than 0.08 mm is greater than 15%. The compaction
specifications should include a specified lift thickness and allowable range of
moisture content above and below optimum. In addition, special attention must be
given to both internal and external drainage.
The peak and residual shearing resistance developed in reinforcement are
dependent on the density of the soil, the effective overburden pressure and the
geometry and surface roughness of the reinforcing elements. Higher densities of
soil can increase the normal stresses acting on reinforcing elements and the
apparent coefficient of friction, at least in those ranges of overburden pressures
where the soil is dilatant (i.e. in which the void ratios are less than the critical
void ratio). Hence, a knowledge of the void ratio, density, shear strength and
state of strain of the soil is important when selecting the apparent friction
coefficient. Obviously the nature of the surface of the reinforcement should also
be considered.
Granular soils compacted to densities that result in volumetric expansion
during shear are ideally suited for use in reinforced earth structures (Vidal,
1969). If these soils are well drained, then effective normal stress transfer
between the strips and soil backfill occurs instantaneously as each lift of backfill
is placed. Furthermore, the increase in shear strength does not lag behind vertical
loading. Granular soils behave as elastic materials within the range of loading
normally associated with reinforced earth structures. Therefore, no post-
construction movements associated with internal yielding or readjustments,
should be anticipated for structures designed at working stress level.
On the other hand, fine-grained materials are not especially suitable for most
reinforced earth structures (Vidal, 1969; McKittrick, 1979). The adhesion
between fill and reinforcement is poor and may be reduced by an increase in
pore-water pressure. Such soils are normally poorly drained and effective stress
transfer is not immediate. Hence a much slower construction schedule or an
unacceptably low factor of safety in the construction phase is necessary.
Moreover, cohesive materials often exhibit elastoplastic or plastic behaviour,
thereby increasing the possibility of post-construction movements. Nonetheless,
such fill can be used successfully in structures like embankments. Furthermore,
Sridharan et al. (1991) showed that cohesive soil with low internal friction
angles can be used as backfill material when layers of sand are used in contact
with the reinforcing strips. The thickness of the layer of sand depends largely on
the surface roughness of the reinforcement and the strength of the bulk backfill
and to a lesser extent on the shape and size of the reinforcement. Generally,
however, a sand layer 15 mm thick is sufficient to increase the interfacial friction
angle to one almost equal that of a completely granular backfill. Hence the pull-
out resistance, pull-out length behaviour of such a sandwiched system is more or
less the same as reinforced earth consisting of granular soil.
7.1.4
Anchored earth
In anchored earth the pull-out resistance is provided by development of bearing
pressures around an anchor at the end of the tensile member which connects it to
the facing unit (Jones et al., 1985). The ultimate load of an anchor corresponds to
the structural failure of the anchor or the anchor pulling through the soil. There
also is a difference between the load-deflection characteristics of an anchor and a
reinforcing strip. Various forms of anchor can be used, one of the most effective
being triangular shaped and formed of mild steel reinforcing bar (Figure 7.8).
PA=2PB
In the second, a triangular wedge of soil is retained within the anchor, shearing
occurs on the top and bottom of the triangular wedge and supplements the
bearing failure at the two leading members of the anchor, hence
PA=PB+2P′F
The value of PB (appropriate to ′=35°), according to Jones et al. (1985), may be
taken as
PB=4kPσv′Bt
where B is the length of the back bar of the anchor which is used to define
anchor size, Kp is the coefficient of passive earth pressure, σv′ is the vertical
effective stress and t is the diameter of anchor bar. A lower bound case for
P′F=Aσv′ tan ′
where A is the plan area of a triangular anchor, v′ is the vertical effective stress
and ′ is the peak angle of shearing resistance measured in a shear box
Figure 7.8 (a) The components and shearing mechanisms of a triangular anchor (after
Jones et al., 1985). There are two alternative mechanisms for shear around a triangular
anchor element. In the first, soil flows around the two leading members and the back
member of the anchor as it is pulled through the soil, so that
(b) Anchored length of construction using full height facings (after Jones et al,
1985)
An anchor formed from round bar offers the smallest surface area which is
advantageous in terms of protecting it against corrosion. The mobilization of
bearing pressures around the anchor, rather than dependence on friction along the
reinforcement, reduces the need for good frictional fills, thereby allowing lower
quality materials to be considered (Temporal et al., 1989). Anchored earth is
cheaper than reinforced earth since the quantity of steel used in the anchors is
about 60% less than that required for reinforcement.
A similar concept has been developed and used in Japan. For example,
Fukuoka and Imamura (1982) described the construction of an anchored earth
retaining wall in which each anchor consisted of a 20 mm diameter steel bar
attached to a concrete plate, 400 mm×400 mm, which was embedded in the soil
backfill. Subsequently Fukuoka (1986) suggested a similar system in which the
tie-bars were attached to sandbags embedded in the backfill during construction
(Figure 7.9).
7.2
IN-SITU REINFORCEMENT TECHNIQUES
There are three principal types of in-situ soil reinforcement techniques used to
stabilize slopes and excavations, namely, soil nailing, reticulated micropiling and
Figure 7.9 Schematic cross-section through reinforced soil wall with multiple tie-bars
attached to bags of sand that serve as anchors. The rods and anchors are embedded in the
embankment during construction (after Fukuoka, 1986)
dowelling. Soil nailing is used to reinforce the ground with small inclusions,
usually steel bars. These are installed horizontally or subhorizontally so that they
improve the shearing resistance of the soil by acting in tension (Bruce and
Jewell, 1986, 1987). The face of the soil mass is stabilized with shotcrete so that
the soil nails and shotcrete represent a resistant unit supporting the unreinforced
soil behind, in a similar way to a gravity-retaining wall.
Reticulated micropiles are steeply inclined in the soil at various angles both
perpendicular and parallel to the face (Lizzi, 1977). Again the aim is similar to
that of soil nailing—that is, to provide a stable block of reinforced soil which
supports the unreinforced ground. In this method the soil is held together by the
multiplicity of reinforcement members acting to resist bending and shearing
forces.
Soil dowelling is used to reduce or halt soil creep on slopes. The slopes treated
by dowelling are much flatter than those in the other two types of in-situ
reinforcement.
Reinforcement provides the largest increase in strength when it is angled
across the potential shear surface so that the reinforcement is loaded in tension.
Hence, where a steep slope is excavated in granular soil the most effective use of
reinforcement is to install it through the face in a more or less horizontal
direction. Therefore, in such situations soil nailing is likely to be more cost
effective than reticulated micropiles. In marginally stable granular or scree
slopes, where stability must be improved but where excavation is not foreseen,
then either soil nailing or reticulated micropiling is acceptable. As can be
inferred from above, in flatter slopes in cohesive soils where stability is governed
by a welldefined shear zone, soil dowels are the most appropriate.
7.2.1
Soil nailing
Soil nailing is a technique whereby in-situ soil is reinforced by the insertion of
steel rods, 20–30 mm in diameter (Gassler and Gudehus, 1981). However,
because of the possible corrosion of steel bars, reinforcement coatings with high
resistance to corrosion are being developed and fibre-glass nails have been used
recently (Gassler, 1992). The technique is employed to enhance the performance
of granular soils and stiff clays. Excavations in soft clay (undrained shear
strength less than 48 kN/m2) are unsuited to soil nailing since its low frictional
resistance would necessitate a very high density of very long nails to ensure
adequate stability. Other soils in which soil nailing is not cost effective include
loose granular soils with blowcounts from the standard penetration test of less
than 10 or relative densities less than 30%; poorly graded soils with coefficients
of uniformity of less than 2 (nailing is not practical because of the necessity of
stabilizing the cut face prior to excavation); and highly plastic clays (plasticity
index greater than 20%) due to excessive creep deformation. It normally is used
to improve the stability of slopes and to provide support in excavations (Elias
and Juran, 1991; Figure 7.10). As far as nail length is concerned, it may be
around 50% of the height of the excavation requiring support. In the case of
slope stabilization, nail length depends on the position of the critical shear plane.
Soil nailing is constructed by staged excavations from ‘top down’
(Figure 7.11). It requires the formation of a series of small cuttings generally 1–
2m high in the soil. Cuts of more than 2m or less than 0.5m are rare in granular
soils. Greater depths have been used in overconsolidated clays. These cuts must
be able to stand unsupported for at least a few hours prior to shotcreting and
nailing. The soil must therefore have some degree of cohesion or cementation
otherwise a pretreatment such as grouting may be required to stabilize the face. A
dewatered face in the excavation is desirable for soil nailing. If groundwater
percolates through the face, the unreinforced soil is likely to fail locally on initial
Figure 7.10 (a) Use of soil nailing in the construction of a cut-and-fill tunnel. (b) Use of
soil nailing to stabilize the portals of a tunnel and adjacent cut slopes (from Bruce and
Jewell, 1987)
excavation, making it impossible to establish a satisfactory coat of shotcrete. As
soil nailing is in intimate contact with the excavated soil surface this minimizes
the disturbance to the ground. Soil nailing can proceed rapidly and the
excavation can be easily shaped. It is a flexible technique that can accommodate
variations in soil conditions and work programmes, as excavation progresses.
A level working bench some 6m wide should be available for the operation.
Usually the length of a single cut is dictated by the area of the face that can be
Figure 7.11 Cut slope stabilization construction sequence
stabilized during a working shift. Any loosened areas on the face should be
removed prior to facing support being applied. Normally the face support must
be placed as early as possible to prevent relaxation or ravelling of the soil. The
support involves pinning a reinforcement mesh to the face and spraying with
shotcrete before drilling the nail holes. The thickness of shotcrete varies from 50
to 150 mm for temporary applications to 150 to 250 mm for permanent projects.
In the ‘Hurpinoise’ system the angle steel reinforcement is often driven before
placing the mesh and shotcrete. A recently introduced system involves inserting
the nails into the soil by means of a compressed air ‘gun’. Under favourable
conditions nails, 38 mm in diameter and up to 6m in length, can be inserted into
the soil at a rate of one every two to three minutes.
A drainage system is required to drain the reinforced zone, and consists of
shallow and deep drains. The shallow drains are tubes which convey water from
immediately behind the facing. Deep drains consist of slotted tubes and are
angled upward at 5–10°. They are usually longer than the nails.
After placement of shotcrete, a series of holes are drilled at, for example, 1.0–
1.5m centres, and are inclined into the soil. The nails are either drilled into the soil
or grouted into predrilled holes. Generally borehole diameters range from 75 to
150 mm for drilled and grouted nails. This usually allows a grout annulus of at
least 20 mm thickness around the reinforcement providing a degree of corrosion
protection. Normally cement grouts are used and are injected at low pressures. In
the jet-bolting system very high pressures (over 20 MPa) are used to inject
cement grout through small apertures at the top of the nail while it is being
installed. This grout lubricates the penetration of the nail and enhances its bond
capacity. Alternatively, nails can be driven into the soil. The rate of installation
can be very high when nails are driven. However, driven nails may be less
suitable in tills and dense cemented soils than grouted nails. Also, care must be
taken in loose, weakly cemented granular soils to ensure that driving does not
cause local destructuring of the soil around the nail which could result in low
values of bond stress. Lastly, the exposed ends of the nails are covered with
shotcrete. The whole process is then repeated for the next layer of soil excavated.
A nailed soil structure must be able to resist the outward thrust from the
unreinforced soil behind without sliding, and must be stable against deep-seated
failure mechanisms. The nails must be installed in a dense enough pattern to
ensure an effective interaction with the soil in the reinforced zone, as well as
being long enough and strong enough to ensure a stable reinforced zone. Because
nails are installed at high density the consequences of one failing is not usually
significant. The nail is placed in the ground unstressed and the reinforcement
forces are mobilized over its length by subsequent deformation of the soil. As
noted above, grouting techniques are usually employed to bond the nail to the
surrounding ground. Where the nails are driven, reinforcement forces are
sustained by the frictional bond between the soil and the nail. Overall
movements needed to mobilize the reinforcement forces are surprisingly small.
Because of the flexibility of a soil-nailing system it can resist seismic loading.
For instance, three soil-nailing projects in California, which were within 33 km of
the epicentre of the Loma Prieta earthquake of 1989, magnitude 7.1, survived
undamaged (Ferworn and Weatherby, 1992). Recent accounts of the design of soil
nailing have been provided by Bridle (1989), Jewell and Pedley (1990a, 1990b),
Bridle and Barr (1990), Juran et al. (1990b) and Pedley et al. (1990).
A new method of soil nailing uses geotextiles, geogrids or geonets to cover the
ground surface (Koerner and Robins, 1986). The geosynthetic material is
reinforced at distinct nodes and anchored to the slope using soil nails at the
nodes (Figure 7.12). When the nails are properly fastened, they pull the
geosynthetic material into the soil, placing it in tension and the constrained soil
in compression.
7.2.2
Reticulated micropiles
Small-diameter (75–250 mm) reticulated micropiles, or as they are sometimes
called root piles, are installed in cased boreholes. In the smaller diameter range
the concrete micropiles are provided with a central reinforcing rod or steel pipe,
while those with larger diameters may be provided with a reinforcing bar-cage
bound with spiral reinforcement. The principal difference between reticulated
micropiles and soil nailing is that, in the former, the reinforcement by the root
Figure 7.12 Schematic cross-section of anchored geosynthetic ‘spider netting’ used with
soil nails to stabilize a slope (after Koerner and Robins, 1986)
piles is strongly influenced by their three-dimensional geometric or root-like
arrangement (Figure 7.13). The holes are formed by a high-speed rotary drill
and, as drilling proceeds, a steel casing is gradually driven into the hole to
prevent its collapse and to limit loosening of the soil. Reinforcement, usually a
single steel bar, is placed in the centre of the casing. The hole is then filled with a
high slump, cement-rich, small aggregate concrete. During withdrawal of the
casing additional concrete is placed by pumping or compressed air pressure. This
produces an intimate contact between concrete and soil, forces penetration of the
concrete into cracks and fissures and, in the case of compressible soils, produces
a pile of larger diameter than the original casing. Reticulated micropiles can be
installed in gravels, sands, silts and clays. Drilling can take place in restricted
areas and does not produce detrimental vibrations.
A cluster of reticulated micropiles, usually containing some installed at a
batter, can be used to form a type of in-situ reinforced earth mass or monolithic
block (Lizzi, 1977). Used in this way these reticulated micropile structures have
applications in slope stabilization (Figure 7.13), retaining structures, and
underground construction (Figure 7.14). Individual piles within a reticulated
micropile structure may be called on to carry tension, compression and flexural
stresses so that the interactions with the included soil are complex. Cantoni et al.
(1989) described the use of reticulated micropiles, together with ground anchors,
to stabilize a sliding slope (Figure 7.15) which was affecting a motorway. Both
the micropiles and anchors were attached to connecting beams.
Figure 7.13 Reticulated micropiles used for slope stabilization (after Lizzi, 1977)
Figure 7.14 Protection of tunnel by reticulated micropiles

7.2.3
Soil dowels
The stabilization of slopes in stiff clay which are undergoing soil creep by using
dowels (Figure 7.16) has been described by Gudehus and Schwarz (1985). They
also outlined their design method. The dowels must be distributed so that the soil
cannot move between them. The most efficient way to improve the shearing
resistance on a weakened shear surface in soil is to use relatively large-diameter
dowels (e.g. 1.5m diameter) with high bending stiffness. Consequently the
Figure 7.15 Stabilization of a sliding slope with reticulated micropiles and ground
anchors attached to connecting beams (after Cantoni et al., 1989)
diameter of a soil dowel is generally far greater than that of a soil nail or
micropile. Dowel action only occurs after a lapse of time that is needed for
displacement to mobilize a sufficiently high lateral force. The dowels are made of
concrete or steel and they transmit the stabilizing force from the substratum to
the creeping soil. The lateral load on the dowels increases linearly up to a
maximum value with displacement relative to the surrounding soil.
7.3
SOIL ANCHORS
Soil anchors can be regarded as a form of tension pile comprising a rod or cable
grouted into a borehole and stressed after the grout has set. In the case of soft
ground excavations, they are used mainly as structural ‘ties’. For example, they
can be used to support diaphragm or sheet-pile walls (Littlejohn, 1990). They can
also be used to stabilize slopes, and to strengthen and consolidate soil.
The capacity or pull-out resistance of a grouted anchor depends on the amount
of skin friction resistance and end-bearing resistance developed by the anchor as
it is loaded (Figure 7.17). In the case of granular soils this resistance also
depends on the depth, density and value of the angle of shearing resistance of the
soil, as well as on the anchor dimensions. The size of the diameter of the
anchorage depends on the type of soil and the method of construction. Cement
grout is used for anchor construction and can limit the size of the diameter of the
anchor in many soils. For instance, only in coarse sands and gravels are cement
Figure 7.16 Dowelling in soil (after Gudehus and Schwarz, 1985)

particles able to permeate the soil, radial penetration of grout forming a large
anchor from a small-diameter borehole (Figure 7.18(a)). Resistance to
withdrawal is governed principally by side shear but an end-bearing component
may be included when calculating the pull-out capacity. Anchors in such ground
are designed to resist safe working loads of 80t (Littlejohn, 1970). The
construction of the anchorage is carried out by drilling and driving a casing (102
mm nominal diameter) to the required depth, lowering a prepared steel cable or
bar and then injecting grout under a nominal pressure (30–1000 kN/m2) as the
casing is withdrawn over the fixed anchor length.
In low-density soils, a larger anchor may be obtained by compaction of the
soil. For example, gravel can be placed in the borehole and a small casing, fitted
with a non-recoverable point is then driven into the gravel, forcing it into the
surrounding soil. The process enhances adhesion and produces a larger, uneven
anchor, even if penetration is limited (Figure 7.18(b)). In addition, the grout also
increases the density and angle of shearing resistance of the soil as well as
forming an anchorage. Safe working loads of up to 30t have been mobilized with
this type of anchor (Littlejohn, 1970).
In stronger soils the ground may be subjected to hydrofracture. A packer is
used in order to control hydrofracture in a lateral direction and so increase the
diameter of the anchorage by radial fissuring (Figure 7.18(c)). Where stage
grouting along a fixed anchor or regrouting are envisaged, a tube-à-manchette
Figure 7.17 A stressed soil anchor. FR=pull-out resistance, RS=skin friction resistance,
Rp=end-bearing resistance, L=length of anchorage, d=diameter of anchorage
can be employed. High-pressure grouted anchors are used mainly in cohesionless
soils, although some success also has been obtained in stiff cohesive deposits.
When permeation, compaction or hydrofracture are not acceptable, tension
piles (Figure 7.18(d)) or underreaming (Figure 7.18(e) and (f)) can be used. With
tension piles, a large-diameter borehole or an extremely long anchor may be
required to obtain a realistic anchor resistance, but in dense, deep sands,
sufficient resistance may be developed by a 100 or 150 mm diameter anchor.
Underreaming in sands and underreaming or ‘belling’ in clays spreads the
anchor load and effectively increases the diameter of the anchorage. Indeed, in
the case of clay anchors, underreaming is normally considered essential because
of reduced adhesion at the clay-anchor interface due to softening of the clay in
contact with drilling water or bleed water from grouting. Because clay is
softened by water, the time taken for drilling, underreaming and grouting
operations should be kept to a minimum. Wherever possible drilling for and
grouting the fixed anchor should be done on the same day. In fact, when working
in stiff-fissured clays a delay of 6 hours could be critical due to the deterioration
of the borehole. An expanding underreamer is used to form a bell or a series of
bells in the fixed anchor zone of the augered borehole, the number being
influenced by the strength of the clay, after which the cable or rod is homed and
grouted. Resistance to withdrawal is dependent primarily on side shear with an
end-bearing component, although for single or widely spaced underreams the
ground restraint may be mobilized chiefly by end bearing. This type of anchor
mobilizes the full undrained shear strength of the clay and can resist safe
working loads of 60t. However, flushing water can penetrate fissures in clay
during underreaming and initially may reduce the value of cohesion by half.
Figure 7.18 Various types of soil anchorages L=length of anchorage, d= diameter of

anchorage
Underreamed anchors must have their bells at least 5–6m beyond any assumed
slip plane.
Underreaming is ideally suited to clays with an undrained shear strength
greater than 90 kN/m2. Some difficulties, such as local collapse or breakdown of
the neck portion between the underreams, are likely to occur when the undrained
shear strength varies between 60 and 70 kN/ m2. At less than 50 kN/m2
undrained shear strength, underreaming becomes virtually impossible. The use
of high-pressure grouted anchors may be tried in such circumstances.
8
Geosynthetic materials
Geosynthetics are relatively thin, flexible polymeric materials. Over the past 25
years there has been a tremendous increase in their use due to the development
of a large range of new materials which possess very different mechanical
properties from those available previously (Giroud, 1986). When these materials
are included in soil they improve its engineering performance and also lower the
cost of construction.
8.1
TYPES OF GEOSYNTHETIC MATERIALS
Two main types of geosynthetics are used in geotechnical engineering: that is,
woven and non-woven material. Usually they are composed of a single man-
made material or a combination of two. The most widely used materials are
polyamide (nylon), polypropylene, polyester (terylene) and polyethylene.
Polyvinylidene chloride (PVC) is also used.
Nylon is a tough material, possessing good all-round properties (Table 8.1). It
is available in an extensive range of fabric forms. However, it may undergo a
slight loss in strength when soaked in water. Polyester generally offers a high
resistance to breaking and tends to be the strongest of the more common
polymers and the least extensible of the materials used in these fabrics. It also
has good acid, abrasion and ultraviolet light resistance. Polyethylene combines
high strength against breakage, good durability and resistance to breakdown by
animals, fungus and chemical attack. It has a low density and it produces a
strong, light fabric. Polypropylene is one of the weaker and more extensible
synthetic fibres, and has a tendency to creep under constant loading.
Polyvinylidene chloride has a high abrasion, good chemical resistance to
chemicals and resists ultraviolet light.
Two main types of woven fabrics are available, namely, those manufactured
from extruded filaments and those made from split film tapes. Extruded
filaments have a circular cross-section and are stronger but more expensive to
produce, weight for weight, than flat types. All woven fabrics are characterized
by high strength and low breaking strains—that is, they possess low
Table 8.1 Mechanical properties, fabric form and resistance to attack of synthetic fibres
used for geotextiles (after Cannon, 1976).
* Note: Durability, long-term filtration properties, particle retention, water/liquid

permeability, thickness and weight are all properties which depend upon the fabric
construction process and have to be considered where appropriate.
GEOSYNTHETIC MATERIALS 175
extensibility. This is not necessarily an advantage in soft ground where more

extensible fabrics may prove more successful. Woven fabrics exhibit a slightly
higher strength at 45° to the warp and weft directions than along them. In
addition, they possess up to double the extensibility at 45°, than parallel to warp-
weft directions. Unfortunately, if a woven fabric is torn, the tear can develop
very rapidly at low applied stresses. However, woven fabrics made from multi-
filament threads use several strands and so offer good levels of resistance to
tearing.
Rankilor (1981) subdivided non-woven fabrics into two types, namely, thin
and thick. The former are commonly referred to as two-dimensional fabrics,
whereas the thicker fabrics are termed three-dimensional fabrics, or felts.
Two-dimensional fabrics have a uniform strength in all directions and possess
a wide range of pore sizes. However, they generally are not as strong as three-
dimensional fabrics and they are more extensible. On the other hand, their two-
dimensional structure means they are not as susceptible to clogging by particles
of soil.
Non-woven fabrics are produced in three ways: by needle punching, by melt
bonding and by resin bonding. In the needle-punching process a layer of
randomly oriented continuous fibres is laid on a felting machine and barbed
needles are driven repeatedly through the fibres. Hence, the fibres are
mechanically interlocked, being held together by physical entanglement
(McGown, 1976). Fabric weights of 300–1000 g/m2 usually are necessary to give
acceptable levels of fabric strength for civil engineering purposes. Generally
needle-punched non-woven fabrics offer a high resistance to tearing as the stress
causes the filaments to move towards the direction of the applied force.
Melt-bonded fabrics consist of fibres which are melt bonded together at cross-
over points. Higher strength is achieved with this type of manufacture at lower
fabric weights (70–300 g/m2) than with needle-punched fabrics. Two methods of
bonding are used, homofil bonding and heterofil bonding. In the former (which
produces spun-bonded fabric) the filaments are composed of a single polymer
type but some of the filaments have different melting characteristics. In the
latter, some of the filaments are composed of two types of polymer that have
different melting points (heterofilaments). Spun-bonded fabrics exhibit a range
of pore sizes in an individual fabric and generally have a lower extensibility than
other melt fabrics. Non-woven materials produced by the melt-bonded process
are generally isotropic and offer a high mechanical resistance. However, when the
filaments have tightened, the resistance to tearing may be low. Also, the rigid
bond between filaments can prevent the diffusion of localized mechanical
deformations.
Usually in resin-bonded fabrics acrylic resin is sprayed onto or impregnated into
a fibrous web. After curing, strong bonds are formed between filaments.
Geogrids and geonets are formed by either punching holes in a thick extruded
sheet followed by unidirectional or bidirectional drawing to form an orthogonal
grid, or by joining thick strands of extruded filaments at required cross-over points.
Polypropylene and high-density polyethylene are used to manufacture geogrids,

which means that they can resist attack by acids, alkalis and salts encountered in
soils. They have a very long life when incorporated with soils and can maintain
their engineering properties for decades even when exposed to sunlight.
Furthermore, the molecular orientation of the polymer which occurs during its
manufacture gives a high directional strength, as well as a high elastic modulus
which ensures mobilization of high tensile strength at low strains and
dramatically reduces creep. Geowebs are formed by weaving thick strands
together.
Geomembranes are non-porous sheets that have very low permeability. The
sheets may be formed of polyethylene, or non-woven polypropylene may be
bonded to polyethylene.
Mats are robust three-dimensional geosynthetics with high tensile strength
which are used to drape over soil slopes. The mat provides an erosion-resistant
cover to the slope.
Geocomposites consist of an inner, more porous material sandwiched between
two fabrics (Anon., 1989(b)). They are used in increasing quantities to replace
conventional drainage systems.
8.2
PROPERTIES OF GEOSYNTHETIC MATERIALS
Synthetic fibres have many excellent qualities and can be individually designed
to cover a wide range of properties. For example, they can possess high strength;
flexibility; ability to elongate without rupture under high loadings and repetitive
stresses; stability in chemicals, groundwaters, effluents and soils; resistance to
fungal attack; and high abrasion resistance (Table 8.1).
Geosynthetics are composed of polymers, and the influence that the polymers
have on behaviour depends very much on the structure of the geosynthetic
produced by the manufacturing process. In the case of woven products in which
the filaments, yarns or tapes of polymer are aligned in the direction of applied
load, the behaviour of the geosynthetic relates closely to that of the basic
polymer. By contrast, where the fibres or filaments of the polymer are randomly
laid during manufacture and linked by bonding or needle punching, the influence
of the polymer on the behaviour of the geosynthetic is very much less. In other
words, the polymer properties have little influence on the properties of the
fabric.
Enclosure of woven geosynthetics within soils has little effect on their load-
strain behaviour, whereas with non-wovens, particularly needle-punched
products, there is a considerable change in their behaviour. This is principally
due to the soil particles locking the fibres into position, so reducing the amount of
straightening and slippage at crossovers and ends. This effect depends on the
type of soil in contact with the geosynthetic, with fine sands and gravels giving
the most pronounced effects.
The critical properties of geosynthetics depend upon the function they are
meant to serve (McGown et al., 1982). For example, pore size and shape are
critical factors when a fabric is being used for filtration. When used as
reinforcement the overall load extension and surface friction properties are
critical. The structural arrangement of many geotextiles is liable to change when
subjected to compressive stresses over their surfaces. Generally, the properties of
woven fabrics are highly anisotropic as compared with those of non-woven
fabrics.
8.2.1
Physical properties
The specific weight of a fabric is a fundamental property, being related to its
strength, deformability and durability.
The rigidity of the structure of a fabric is as important as its complexity as far
as filtration is concerned. For instance, a fabric that is open in the unloaded and
loaded states gives a measure of the rigidity of its structure. A thin simple fabric
structure is suitable for most cases of static filtration, whereas a thick complex
fabric is required for plane drainage.
The pore size distribution of fabrics does not provide a direct measure of their
particle-retaining capacity. Additional factors such as the particle size
distribution of the soil, the flow rate and flow pattern of groundwater and particle
size distribution of the downstream soil, also influence the particle-retention
capabilities of a fabric in a particular application. Nonetheless, the pore size
distribution of an unloaded fabric offers a means of comparing the absolute pore
size and size distribution of fabrics (Figure 8.1). The range of porosity of non-
woven fabrics is much greater than that of woven fabrics. A non-woven fabric
may possess a high porosity; for example, it may exceed 90% in the unloaded state
or be 70–80% when loaded. The nature of the changes in the pore space of
geotextiles vary with the level of confinement and the character of the material
with which they are in contact. Changes in pore space alter the particle-retention
capacity and permeability of the fabric. The permeability of a fabric, depending
on type of fabric used and structural arrangement, may vary by more than a factor
of 10.
The hydraulic conductivities along and across fabric are significant
characteristics in many applications. In order to gauge the influence of changes
in fabric structure during loading, values of fabric permeability should be known
for the unloaded and loaded conditions.
8.2.2
Mechanical properties
The stress-strain behaviour of geosynthetics exhibits an extremely wide range—
for instance, some woven fabrics have strengths, moduli, break strains and creep
Figure 8.1 Typical fabric pore size distributions (after McGown, 1976)
properties which approach that of steel mesh (Figure 8.2). These woven fabrics,
therefore, can be used instead of steel or aluminium strips in earth reinforcement.
On the other hand, some non-wovens have relatively low strengths with breaking
strains approaching 150% and high creep tendencies (Figure 8.3). The stress-
strain characteristics of fabrics provide an indication of their relative toughness.
In situations where large strains have to be accepted, extensible non-woven
fabrics prove suitable. Conversely, the use of a stiffer woven fabric is necessary
when strains in the soil must be limited.
Generally, geosynthetics are more extensible than the soils in which they are
placed in that they need a greater amount of tensile strain to cause rupture than
soils require to achieve peak strength (McGown et al., 1982). Hence, the tensile
resistance that can be utilized in geosynthetics is very much less than their
rupture strength. In fact, in most cases the strength that can be utilized is limited
by the permissible strains in the soil, which rarely exceed 10% and in many
applications are very much less. Consequently, as far as soil reinforcement
applications are concerned, it is the load-strain-time behaviour of geosynthetics
up to 10% strain that is important.
Figure 8.2 Stress-strain curves for various filaments (from Hoare, 1987)
The peak tensile strength and strain measured under plane strain conditions
offers an indication of the maximum load capacity of a fabric and the strain
required to achieve this. Woven fabrics, on a weight for weight basis, afford
much higher resistance to breaking than non-woven types. The latter generally
are characterized by relatively low strength and high failure elongation. In
general, the percentage loss in strength is much less for non-woven than for
woven fabrics. However, the higher initial strength of woven fabrics must be
borne in mind. The structural arrangement of a fabric may give rise to various
degrees of anisotropy in terms of strength and extensibility. Changes in the
internal structural mobility, interfibre friction and the surface texture of
geotextiles are likely to occur according to the level and nature of confinement.
In this way the load-extension and surface friction characteristics of a fabric are
altered. In fact, Ingold (1991) referred to a transition point at which
the mechanism controlling tensile failure underwent a change, the time at which
this occurred being affected by polymer type, environmental conditions and any
damage caused by construction operations. Because the tensile strengths of
different geosynthetic products differ and are affected to differing degrees in
different environments, the new British code of practice for reinforced soils
(Anon., 1991) has adopted the concept of partial factors of safety, using a partial
load factor and a partial material factor to obtain a margin of safety against
failure.
In certain situations, especially in extensible strain reinforcement situations
where the inclusions redistribute the strains within the soil mass, loads are
applied unevenly within the fabric. The ‘grab’ tensile test can be used to assess
localized loading and takes account of any contribution made by the fabric that is
not in the direct stress path between loaded points. This contribution can be large
Figure 8.3 Typical plane strain data for different types of geotextiles, normalized to 300 g/
m3 fabric weight (after McGown, 1976)
for some materials, particularly non-wovens with highly random arrangements of

fibres. Typical values for fabrics are given in Table 8.2.
The bursting load of fabric and extension to burst are important properties.
Resistance to small deformations that are likely to occur on site can be measured
with a Mullen burst test. Values of bursting load obtained by such tests are given
in Table 8.3. Tear propagation, rather
Table 8.2 Typical grab tensile characteristics of some fabric types used in civil
engineering (normalised to 200g/m2) (after McKeand and Sissons, 1978)
Breaking load Extension to Load at 5% Rupture energy
(N) break (%) extn. (N) (Nm)
Melt-bonded: 900–1400 70–125 110–240 50–90
Melded
Other 1000–1200 50–60 250–300 40–50
Needle-punched 750–950 65–150 10–20 20–65
Resin-bonded 650–800 50–65 100–250 20–25
Woven tape 800–1000 10–25 300–350 10–25
than tear initiation, is another important property. In some applications such as

river bank and coastal protection, this may be one of the major requirements of
the fabrics used. Also they should not be affected significantly by punching
loads.
When loads have to be sustained continuously—as, for example, in earth
reinforcement—the creep-relaxation properties of the fabrics are of major
consequence. Creep is a function not only of the type of material, the production
method and form of reinforcement, but also of the level of stress imposed, the
length of time involved, environmental attack and the operational temperature.
Creep can result in rupture at lower than the recognized breaking stress and in
order to apply satisfactory factors of safety, the load placed on fabric should be
limited to less than 50% of the measured breaking load. Generally, the load at
which tensile rupture occurs, at a given temperature, decreases with time. Creep
coefficients for polyester are lower for lower levels of loading, though this is not
the case for nylon yarn. Creep in polypropylene fibre is generally of a higher
order. When a fabric is subjected to tension over a
Table 8.3 Typical values of Mullen burst for some fabric types used in civil engineering
(normalised to 200g/m2) (after McKeand and Sissons, 1978)
Bursting load (N/cm2) Height of dome at burst (cm)
Melt-bonded: Melded 100–200 1.4–2.1
Other 150–200 1.4–1.6
Needle-punched 100–210 1.1–1.9
Resin-bonded 150–270 0.9–1.1
Woven tape 350–380 1.0–1.2
period of time, stress decay measurements provide an indication of creep

behaviour. Such measurements show that needle-punched non-woven fabrics
exhibit better stress decay characteristics than melt-bonded materials under static
loading. The situation is dramatically reversed under cyclic loading (Figure 8.4).
The transfer of load from soil to fabric and between overlapped layers of
fabric is dependent on friction. Hence, the coefficients of soil-fabric friction and
fabric-fabric friction are important properties. As the fabric is subject to
frictional contacts, the abrasion resistance of the fabric is also important.
8.2.3
Environmental properties
Corrosion of geosynthetics can take place due to the presence of certain
chemicals in substantial amounts in aggressive acid or alkaline soils. For
example, the presence of sulphuric acid, together with the activity of sulphur
bacteria, can bring about deterioration of nylon-based products. Polyamides are
attacked by strong acids and polyesters are susceptible to strong alkalis, but both
have excellent resistance over the pH range (3–10) likely to be encountered in
most soils. Oxidation can cause degradation of nylon, polyesters and
Figure 8.4 Stress decay on static and cyclic loading of non-woven fabrics (after McKeand
and Sissons, 1978)
polyolefines. Water is absorbed by and diffused through most polymers at

ambient temperatures and humidities. Indeed, in composite materials the
possibility of preferential diffusion of water exists, diffusion occurring due to
capillary action along, for example, interlaminar planes.
Some micro-organisms present in soils are detrimental to synthetic products.
Microbial activity is most significant in humid tropical areas.
It appears that the tensile strength of some types of nylon deteriorates with time
after incorporation in soil. In some cases the material has been found to become
brittle, rupture then occurring without any appreciable elongation. This may be
attributable to hydrolysis. The strength of polyester, polythene and
polypropylene undergo negligible or insignificant changes after burial.
During construction, exposure of geotextiles to some direct sunlight and other
weather effects is unavoidable, the effects of exposure being especially
important in hot climates. Hence, a factor that must be considered when
geotextiles are used in construction is their degradation when exposed to weather
and ultraviolet light. The effects of weathering on geotextiles depends on their
geomediaal location. For example, Brand and Pang (1991) quoted tests carried
out in Florida, North Carolina and Arizona on the one hand, and Sweden and
England on the other. In the North American tests some geotextiles subjected to
exposure had a half-life (time to 50% strength loss) of less than 8 weeks, while
the half-life of others exceeded 32 weeks. The woven polypropylene geotextile
exposed in England and Sweden underwent strength losses of 14–22% in a year
compared with losses of 2–5% per annum for buried samples. The prediction of
the performance of a geotextile upon exposure should be based on expected sun-

hours; the energy of sunlight falling on the geotextile; temperature, humidity and
atmosphere pollution; and the temperature of the geotextile. Moreover, different
polymers have different sensitivities (e.g. polyester is sensitive to wavelengths in
the region of 325 nm, polyethylene to 300 nm and polypropylene to 370 nm).
The weakening of synthetic polymers by ultraviolent light can be partially
compensated for by addition of an UV stabilizing agent (Cooke and Rebenfeld,
1988).
A series of natural exposure tests carried out in Hong Kong by Brand and
Pang (1991) showed that all the geotextiles became brittle with a corresponding
loss in tensile strength. The loss of strength in the first month was less than 16%,
with some geotextiles performing better than others. The long-term performance
of the geotextiles varied considerably, with the half-life ranging from 3 to 9
months. Some geotextiles virtually lost their strength in 6 months. The rate of
strength loss generally increased with radiation intensity. Brand and Pang found
that polypropylene geotextiles without UV stabilizers offered the lowest
resistance to weathering, while needle-punched polyester geotextiles without UV
stabilizers performed in a similar way to the UV stabilized polypropylene needle-
punched geotextiles. Needle-punched geotextiles (≥ 1.45 mm thick) gave better
long-term performance than the thinner and more open-structured woven
geotextiles, irrespective of polymer type. In the case of polypropylene
geotextiles, needle-punched fabrics had better resistance to degradation than the
thinner heat-bonded geotextiles. Accordingly, Brand and Pang recommended
that rolls of geotextiles should remain covered until they are used and that they
then should be buried as soon as possible.
8.2.4
Properties characteristic of geogrids
When placed in a soil the ribs on a geogrid which run transverse to the direction
of primary loading provide a series of anchors (Jewell et al., 1984(a)).
Consequently, stress is transferred to the grid not just by surface friction, as with
strip reinforcement, but also by interaction. This provides a very effective means
of stress transfer which mobilizes the maximum benefit from the grid
reinforcement and minimizes anchorage lengths. The open structure of a geogrid
interlocks and interacts with the soil to provide a high resistance to sliding.
In many reinforcement applications, high modulus geogrids are better suited
than low moduli geosynthetics since they mobilize reinforcement properties at
elongations that are more compatible with working strains in soils. Typical
compacted soils have maximum working strains around 1–4% and at elongations
in this range only very high modulus geogrids can mobilize significant tensile
forces.
8.3
FUNCTIONS OF GEOSYNTHETICS
Geosynthetics perform a number of basic functions, namely, separation
(segregation of two layers of solid particles); cushioning (ability to absorb
impact or abrasion from impinging materials); filtration (passage of liquid/gas
across an interface with retention of solids); reinforcement (redistribution of the
stress-strain pattern in a load-carrying system); and, in the case of thicker
fabrics, drainage in the plane of the fabric (McGown, 1976; Fluet, 1988). In
some applications one function may be dominant. However, many applications
of a fabric involve a combination of more than one of the basic functions.
Geomembranes serve an isolation or protection function.
An example of the separation function is the separation of the aggregate of a
roadbed from the subgrade below. The geotextile prevents the loss of aggregate
to the subgrade while, at the same time, preventing the subgrade material from
being intruded into the aggregate. Thick non-woven fabrics can perform a
cushion function, protecting one material from damage by another. An example
is provided by the use of non-woven geotextiles in liners to
protect geomembranes from overlying gravel. In the filtration function a
geotextile retains soil while allowing the passage of water. This occurs when
filters are used for erosion control as well as drainage. The purpose of a
geosynthetic material when used for reinforcement is to add tensile strength to
the soil and thus construct an earth structure with adequate compressive strength
(derived from the soil) and tensile strength. Geogrids provide higher low strain
elastic moduli than other commercially available geosynthetics, while specialized
woven geotextiles provide the highest tensile strengths. Lateral drainage or
transmission of fluid through the plane of the geosynthetic requires a material of
adequate thickness and permeability. In addition, the material must possess
sufficient dimensional stability to retain its thickness under pressure.
Geomembranes, because of their impermeability, can be used to isolate and
retain fluids. For example, geomembranes may be used to provide isolation (i.e.
waterproofing) in tunnel construction or to protect foundations in aggressive
soils.
The maximum pore size of a fabric represents the largest particle size that can
pass through a fabric without tearing it. However, in most situations this particle
size is retained. Thus the particle-retention capacity and pore size distribution of
fabrics are not the same.
The particle-retention capacity of a fabric is not a constant. It depends on the
characteristics of the materials in contact with it as well as the hydraulic
conditions of the soil in which it is incorporated. Where nonreversing flow
occurs, the particle size distribution of the downstream soil can reduce the
effective size of the pore opening. In addition, if the upstream soil is well graded,
bridging of particles may occur across the pore openings which appreciably
reduces the likelihood of many particle sizes moving through the fabric. The
higher the tortuosity of the pores within the fabric, the more effective is the
retention capacity. In reversing flow conditions the bridging effect may not
develop and particles may be moved in and out of the fabric. This tends to
increase the size of particle migration over that in non-reversing flow conditions.
As the pore size distribution of a fabric may be altered significantly when
loaded, it is necessary when considering the relative usefulness of fabrics
performing as separators, filters, drains or reinforcements to test the fabrics in the
soil (McGown et al., 1978). As far as separation is concerned, most fabrics
perform satisfactorily under non-reversing and low hydrodynamic reversing flow
conditions. Where the dynamic effects are significant the range of applicability
of fabrics can be appreciably increased by using them in conjunction with a layer
of filter sand.
Filtration and drainage-in-the-plane functions occur over a range of flow
conditions, including steady state and dissipating head, non-reversing flow and
variable head-reversing flow. Hoare (1978) noted that thin fabrics are adequate
for non-reversing flow conditions while thick compressible fabrics can be used
for hydrodynamic reversing flow conditions. Alternatively, a thin fabric together
with some granular material can be used to replace one layer of a multi-layer
filter.
Small particles of soil immediately in contact with a fabric are the first to
move through it, creating voids that may be occupied or bridged by other soil
particles. Migration and re-orientation of soil particles continues and eventually
creates an upstream filter within the soil (see Figure 4.1). The development of
this upstream filter depends on the size and number of particles that are involved
in migration and the hydraulic conductivity of the screen. The effectiveness of
the screen, in turn, is governed by the structure of the fabric, the properties of the
upstream soil, the drainage media grading curve and the fluid flow conditions
(McGown et al., 1978).
Woven fabrics and most melt resin-bonded non-woven fabrics do not possess
sufficient hydraulic conductivity in the plane of the fabric to act as drains.
However, the thicker needle-punched non-wovens, as a consequence of their
bulk, offer some flow capacity. A drain is formed by enclosing granular filter
material in fabric (see Chapter 4). Fabric-wrapped drains provide greater
consistency in filter design, installation and performance than conventional
drains. Purpose-made prefabricated geotextiles which contain a core of channels
enclosed by fabric on either side are now used for drainage purposes. The use of
purpose-made drainage fabrics eliminates the use of granular filler material, an
important consideration when suitable granular material is not readily available.
The purpose of placing tension-resistant members in soil structures is to
increase the resistance of these structures to shear failure. The tensile strength of
a fabric affords tensile strength to an earth mass (Hoare, 1978), the tensile
strength of the fabric being used to resist forces parallel to the plane of the fabric
as in earth-reinforced walls. In addition, the shearing resistance of the soil is
increased. The basic operational mechanism is strain controlled (Juran et al., 1990
(a)). In fact, tension can only develop in reinforcing members when incremental
tensile strains are developed in the soil and in the planes of the reinforcing
members. Therefore, no matter what the level of stress, if tensile strains cannot
occur within the soil and in the planes of the reinforcements, then tension will not
be developed in the reinforcements. The inclusion of fabric within an earth mass
redistributes stress in each membrane. Not only are stress patterns altered when
sheet fabrics are used for reinforcement, but strain directions also are changed
substantially. When strips of webbing are used in a reinforced soil system, they
more or less act as passive tie-backs. They are attached to the facing units.
As mentioned above, most non-woven fabrics have elastic moduli much less
than that of the soil, and can strain far beyond the peak strain of the soil.
Consequently, they have a very low load take-off capacity in a fabric-reinforced
system. Nevertheless, the presence of such fabrics in soil enhances its load-
carrying capacity. Where strains in a system must be restricted, then
geosynthetics with the highest values of strength and deformation moduli, and
the lowest break strains, should be used. Conversely, in systems where strain can
be increased significantly, the inclusion of fabrics can lead to more ductile, less
brittle system behaviour. This, in turn, can give rise to greater load-bearing
capacity. A good example of application of the former is in reinforced earth
retaining walls and of the latter is in granular soil embankments over highly
compressible soils.
8.4
USES OF GEOSYNTHETIC MATERIALS
The improvement in the performance of a pavement attributable to the inclusion
of fabric was one of the first ways in which geotextiles were employed in civil
engineering and remains the most important use (Anon., 1990(b)). The
improvement comes mainly from the separation and reinforcing functions, and
can be assessed in terms of either an improved system performance (e.g.
reduction in deformation or increase in traffic passes before failure) or reduced
aggregate thickness requirements (where reductions of the order 25–50% are
feasible for low-strength subgrade conditions with suitable geosynthetics).
The most frequent role of fabric in road construction is as a separator between
the sub-base and subgrade. This prevents the subgrade material from intruding into
the sub-base due to repeated traffic loading and so increases the bearing capacity
of the system. The savings in sub-base materials, which would otherwise be lost
due to mixing with the subgrade, can sometimes cover the cost of the fabric. The
range of gradings or materials that can be used as sub-bases with fabrics is
normally greater than when fabrics are not used. Nevertheless, the sub-base
materials preferably should be angular, compactible and sufficiently well graded
to provide a good riding surface.
The position of a fabric in a system is critical and the use of two discreet
layers can afford much greater improvement than just twice that of a single
layer. McGown (1976) developed a series of fabric-included designs which

suggest modes of fabric inclusion that give increasing degrees of improvement,
depending on subgrade and sub-base types.
If a fabric is to increase the bearing capacity of a subsoil or pavement
significantly, then large deformations of the soil-fabric system generally must be
accepted as the fabric has no bending stiffness, is relatively extensible, is usually
laid horizontally and is normally restrained from extending laterally. Thus,
considerable vertical movement is required to provide the necessary stretching to
induce the tension that affords vertical load-carrying capacity to the fabric.
Fabrics, therefore, are likely to be of most use when included within low-density
sands and very soft clays. Although such large deformations may be acceptable
for access and haul roads they are not acceptable for most permanent pavements.
In this case the geosynthetic at the sub-base/subgrade interface should not be
subjected to mechanical stress or abrasion. When fabric is used in temporary or
permanent road construction it helps redistribute the load above any local soft
spots that occur in a subgrade. In other words, the fabric deforms locally and
progressively redistributes load to the surrounding areas thereby limiting local
deflections. As a result, the extent of the severity of local pavement failure and
differential settlement is reduced.
When fabric is used for unbound temporary access or haul roads, it is placed
on the subgrade and then overlain with aggregate (Figure 8.5). Ground
penetration by aggregate is minimal and the extensibility of the fabric allows it to
conform to the surface over which it is laid. The fabric also prevents sub-base
deterioration by reducing pumping, whereby fines become suspended in water
and are transported into the sub-base by the action of wheel loading. The
incorporation of fabric at the sub-base/subgrade interface constrains the sub-base
and therefore reduces the development of ruts in the road. The subgrade material
undergoes a long-term loss of moisture in its upper layers, which further
increases its strength.
When constructing roads, railways or runways over variable subgrades,
geogrids can be used within a granular capping layer. To achieve maximum
reinforcing effect, the vertical spacing between grid layers should not be greater
than 500 mm. Multiple layer systems are possible since, by interlocking with sub-
base particles, the grid does not create a weak shear plane.
The use of geogrid reinforcement in road construction helps restrain lateral
expansive movements at the base of the aggregate under wheel loading. This
gives rise to an improved load redistribution in the sub-base which, in turn,
means a reduced pressure on the subgrade. In addition, the cyclic strains in the
pavement are reduced. The stiff load-bearing platform created by the interlocking
of granular fill with geogrids is utilized effectively in the construction of roads
over weak soil. Reduction in the required aggregate thickness of 30–50% may be
achieved. Geogrids have been used to construct access roads across peat bogs,
the geogrid enabling the roads to be ‘floated’ over the surface.
Figure 8.5 (a) Loss of sub-base material due to a reduction in subgrade strength; and (b)
use of geotextile to solve the problem
In arid regions geomembrane can be used as capillary breaks to stop the

upward movement of salts where they would destroy the road surface.
Geomembranes also can be used to prevent the formation of ice lenses in
permafrost regions. The membrane must be located below the frost line and
above the water table.
Where there is a likelihood of uplift pressure disturbing a road constructed
below the piezometric level, it is important to install a horizontal drainage
blanket. This intercepts the rising water and conveys it laterally to drains at the
side of the road. A geocomposite can be used for effective horizontal drainage.
Problems may arise when sub-bases are used that are sensitive to moisture
changes—that is, they swell, shrink or degrade. In such instances it is best to
envelop the sub-base, or excavate, replace and compact the upper layers of sub-
base in an envelope of impermeable membrane. A membrane of limited
permeability is acceptable provided that any changes in moisture content in the
encapsulated soil does not result in a critical loss of strength during periods of
heavy rainfall or flooding, such as experienced in semi-arid regions. Soil that is
on the dry side of optimum moisture content is favoured for a membrane-
encapsulated soil layer (MESL), since higher strengths can be obtained,
especially in soil with a high percentage of fines. The MESL brings about a better
distribution of stress in a road pavement than if a membrane alone was used, and
has been used for rapidly constructed road and airfield pavements (Lawson and
Ingles, 1982).
Geosynthetic material with high tensile strength contributes to the load-
carrying capacity of soil that is poor in tension, and is used as reinforcement in
the construction of earth-reinforced walls and embankments (Ingold, 1984(b);
Leshchinsky and Boedeker, 1989). The reinforcement develops tensile forces
which increase the shearing resistance of the soil and improve stability. The
effectiveness of the reinforcement is principally a function of its strength,
stiffness, location within the soil mass and its ability to bond with soil. The force
in the reinforcement may be resolved perpendicular and parallel to the inclination
of the failure surface. The component of the force perpendicular to the failure
surface acts with the angle of friction of the soil to increase the frictional
resistance. The component parallel to the failure surface resists the shear forces
along the surface.
A wide variety of fill materials and facings can be used with geogrids in the
construction of reinforced soil walls. Horizontally laid layers of geogrid provide
the structural stability while the facing contains the fill and provides a suitable
aesthetic appearance. Reinforced soil structures have a high tolerance of
differential settlement thereby avoiding the requirements for expensive
foundations. Wall facings can consist of concrete panels, timber, brick, stone or
gabions. Alternatively, soft facings can be formed by wrapping the grid layers up
the face around successive lifts. Such faces can be detailed to provide vegetation.
In reinforced earth structures, there are three methods of construction which
accommodate settlements that occur within the soil mass: namely, the concertina
method, the telescopic method and the sliding method. In the concertina method
the face is formed by wrapping the reinforcement (geogrid) around the
compacted soil layer so that successive layers of fill are completely enclosed
(Figure 8.6). The exposed geosynthetic material is covered with asphalt
emulsion, shotcrete or hydroseeded soil for long-term protection against
ultraviolet light and weathering. Deformation of the face accommodates the
settlement. In the telescopic method a short section of geogrid is cast into the
facing panels and joined to the reinforcing grid by a dowel bar. Construction
settlements within the soil mass in the telescopic method are accommodated by
the facing panels closing up. A gap is left between the panels when they are
attached to the reinforcing elements. Settlement is accommodated in the sliding
Figure 8.6 The concertina method of construction of retaining walls involves forming the
face by returning the reinforcement over the compacted layer
method by allowing some freedom of the reinforcement in the soil to slide down
the facing while remaining connected to it.
Rising land prices, the scarcity of good quality fill and the need to widen
existing highways provide incentives to steepen slopes and to utilize marginal
fills. In such cases the reinforcement is laid directly on the surface of a layer of
compacted fill and covered with the succeeding layer of fill. Geogrids may be
used to steepen embankment slopes up to 90°. A face support is required for slope
angles between 45 and 90°. The geogrids can be wrapped around successive lifts
of fill, removing the need for rigid facings. Generally geogrid-reinforced slopes
shallower than 45° do not require the wrap-around detail to provide face
stability. In such instances a mat can be laid over the face, prior to placement of
topsoil, to resist erosion and accelerate and promote the growth of vegetation
(Figure 8.7).
Geogrids or geomats can be used in the construction of embankments over
poor ground without the need to excavate the ground and substitute granular fill.
They can allow acceleration of fill placement often in conjunction with vertical
Figure 8.7 Reinforced soil embankment
band drains. Layers of geogrid or geowebs can be used at the base of the
embankment to intersect potential deep failure surfaces. Rowe et al. (1984)
described the use of strong geotextiles to help construct an embankment over
peat deposits. Geogrids can also be used to encapsulate a drainage layer of
granular material at the base of the embankment. Both of these methods help
reduce and regulate differential settlement. A geocell mattress can also be
constructed at the base of an embankment that is to be constructed on soft soil.
The cells are generally about 1m high and are filled with granular material
(Figure 8.8). This also acts as a drainage layer. The mattress intersects potential
failure planes and its rigidity forces them deeper into firmer soil. The rough
interface at the base of the mattress
Figure 8.8 Geocell mattress beneath embankment

ensures mobilization of the maximum shear capacity of the foundation soil and
significantly increases stability. Differential settlement and lateral spread are
minimized.
Traditionally, slips have been reinstated by excavation and disposal of the
failed soil and substitution with a free-draining granular fill. Re-use of the
slipped soil, reinforced with geogrids to intersect potential failure surfaces, has
enabled repairs to be carried out at a quarter of the traditional cost. Geogrids are
usually installed by excavating the failed slope in sections. Excavation is
continued beyond the slip surface to provide sufficient anchorage length for the
layers of geogrid reinforcement. The removed material is then used to reinstate
the slope. Horizontal layers of reinforcement are incorporated at appropriate
vertical spacings to increase the factor of safety and prevent further slip failure.
The procedure is repeated until the whole failed slope has been repaired in a
series of sections. Granular drainage layers may be incorporated within the repair
to further enhance stability. The addition of lime aids handling of particularly wet
clay without affecting the performance of the geogrid.
Geosynthetics can be used when reclaiming poor-quality low-lying land. The
geosynthetic is rolled directly over the ground surface if a permeable fill, such as
sand, is to be laid above. However, if the only material available is impermeable,
then a drainage blanket should be positioned beneath the fabric to convey away
the groundwater. Two or more layers of fabric, each pair separated by a thin
layer of fill, can be used on very soft ground that has a high moisture content.
Improvements of soft ground with geogrids and geomeshes have been described
by Kamauchi and Kitamori (1985) and Watari et al. (1986).
Geotextiles also have been used to reclaim mine or quarry tailings ponds. The
whole surface is covered with a strong permeable membrane which is then
overlain by a thin layer of compacted granular fill. Broms and Shirlaw (1987)
described the use of geotextile to increase the bearing capacity of very soft
tailings in waste ponds so that a high fill, required for preloading, could be
constructed. The fabric was first placed over the area to be stabilized, after which
berms were placed around the perimeter of the area to hold the fabric in place.
Narrow berms were then constructed across the fabric and then widened until the
whole area was covered.
The use of geosynthetics as a filter media is next in importance to their use in
pavement construction. The performance of a filter drain is governed by the
properties of the soil in which it is incorporated, the hydraulic flow conditions
and by the properties of the fabric (Ingold, 1984(a)).
A fabric filter brings about a ‘self-induced’ filter action within the soil
immediately around it (Figure 4.1). In other words, some fines in the soil
Table 8.4 Criteria for preventing soil piping in one-way flow situations (after Giroud,
1982)
Relative density (DR) 1<CU<3 CU>3
Loose (DR<35%) O95<(CU)(D50 )# O95<(9/CU)D50
Intermediate (35% O95<1.5(CU)(D50) O95<(13.5/CU)D50
<DR<65%)
Dense (DR>65%) O95<2(CU)(D50) O95<(18/CU)D50
where e is the naturally occurring void ratio, emax is the maximum void ratio and emin is
the minimum void ratio.
#O =apparent opening size of geosynthetic; unfortunately this is not easy to determine
95
D50=particle size corresponding to 50% finer than maximum size
CU=D60/D10=coefficient of uniformity
next to the fabric can pass through it and so be transported away while most
remain in place. This process of filter development continues until a new
equilibrium condition is established. From then on only clean water passes
through the filter. The filter criteria of geotextiles can be related to the particle
size characteristics of the soil concerned. For example, Giroud (1982) suggested
a set of criteria to prevent piping (i.e. soil passing through the geotextile), which
are given in Table 8.4.
Fabrics have been used to form various types of drains such as linear drains
where fabric encapsulates granular material in a trench, drainage blankets
beneath fills, or fin drains (see Chapter 4). In the latter case the geocomposite
consists of a sandwich of thin, high-porosity filter fabric enclosing a core able to
collect water. This runs to a perforated unplasticized PVC pipe around which it
is wrapped (see Figure 4.3). The core is formed from an extruded polyethylene
mesh sheet. Geocomposite is used to drain both sides of French drains, and in
Figure 8.9 Double liner system containing two flexible membrane liners (FML) for
disposal of hazardous waste
cut-off drainage where the core is sheathed on one side by filter fabric and
impermeable geomembrane on the other.
Geomembranes are incorporated into clay blankets to form composite liners
for waste disposal sites. Although resistant to many chemicals, some
geomembranes are susceptible to degradation by organic solvents. Hence, the
United States Environmental Protection Agency recommends the use of double
liner systems (Figure 8.9).
Geosynthetics are used for erosion control, for example, to retain soil
undergoing erosion, or in the construction of river and marine defence works.
Geomats, as noted, are three-dimensional materials with high tensile strength and
excellent drape qualities. They are used to protect slopes, the mat stabilizing the
surface layer of soil, and when filled with soil help establish a vegetative cover
on slopes. Mats are draped over the slope requiring protection and should
maintain intimate contact with the soil beneath. Any existing vegetation should
be cropped to ground level or removed, and bumps and hollows levelled out. The
mats are pegged to the soil, the pegs being inserted at the edges of the mats. The
pegs vary in length between 100 and 300 mm. On steeper slopes the mats should
be buried in shallow V-shaped trenches both at the toe and just beyond the crest
of the slope. Any seeding or application of fertilizer to the soil obviously should
be carried out prior to the mats being laid.
Biodegradable geotextiles—that is, prefabricated natural materials —are used
to accelerate the establishment of vegetation on newly formed slopes, thereby
increasing their erosion resistance and stability against shallow-seated slope
failure (Barker, 1986). These geotextiles take the form of open mesh jute (or
hessian) netting, and three-dimensional quilted mats consisting of coir (coconut
fibre) and straw mixtures, or poplar or pine shavings. Biodegradable mats are
essentially reinforced mulches, shielding soil from erosion by rain and insulating
it from extremes of temperature and moisture. They can retain more than their
own weight of water, thereby providing resistance to dry spells. In addition,

biodegradable geotextiles can contain seeded mixes (e.g. grass or tree seeds
mixed with cotton waste). Such materials protect the slope, retain soil moisture,
enhance the humus content of the soil when they degrade and offer
reinforcement to the turf that forms.
Geosynthetics used in river defence works usually require high strength.
Consideration has to be given to the velocity of river water during placing, as
well as to the type of soil being protected. Above water the fabric is laid over the
soil; however, fabric placed below water usually needs to be weighed down. The
rip-rap overlying the fabric should be placed at the toe first and then built up the
slope. Webbing can be used directly for river defences without stone cover.
Flood bunds can be reinforced with geogrids.
A fabric may be placed behind a heavy sea defence structure to act as a
backing separator. In such a case it is necessary to dissipate uplift pressures and
relieve groundwater flow without clogging occurring. A fabric beneath heavy
armour also acts as a filter and may have to dissipate energy as well as withstand
direct wave pressure and wave-induced movement of stones. Geogrids can also
be used to form gabions and mattresses for cliff protection, sea walls or
permeable groynes.
9
Grouts and grouting
9.1
INTRODUCTION
Grouting usually refers to the injection of suspensions, solutions and emulsions
into pores in soils to improve their geotechnical characteristics. Hence
permeation grouting is the commonest type of grouting used in construction.
However, grout may be used to displace soils as in compaction and claquage
(hydrofracture) grouting or to replace soils as in jet grouting. Grouting is widely
used to reduce the permeability and/ or to increase the strength of soil. If
grouting is to be effective, it must reliably penetrate the ground to the required
distance and arrive in a satisfactory state.
Grouts can be grouped into two basic categories: suspension or particulate
grouts (Bingham fluids); and solution or non-particulate grouts (Newtonian
fluids). Particulate grouts consist of cement-water, clay-water or cement-clay-
water mixes. The most common classes of chemical grouts are silicates, lignins,
resins, acrylamides and urethanes (Karol, 1968; Chi and Yang, 1985). The
silicate grouts are the most widely used chemical grouts (over 90% of present
use).
Grout used to reduce the permeability of the ground must be able to develop
sufficient strength to withstand the hydraulic gradient imposed. Normally a
cement or clay-cement grout is used in coarser soils and clay-chemical or
chemical grouts are used in finer grained soils (Figure 9.1). Microfine cement
has been used to grout fine sands (Arenzana et al., 1989). In the case of chemical
grouts, especially on large jobs, cheaper high-viscosity grouts may be used to fill
the larger voids and more costly low-viscosity grouts to fill voids too small to be
penetrated by the initial grouting. While it is not possible to state with any
certainty what will be the result of any given grouting programme, it is fairly
certain that a given degree of watertightness can be achieved if sufficient
grouting is performed. In general terms a cut-off with a permeability of 10−6 m/s
in alluvium is attainable, although in individual cases an improvement of one or
more orders of magnitude above this figure has been claimed.
Figure 9.1 Soil size limitations on grout permeation (after Mitchell, 1970, with
permission of ASCE)
Grouting can be used to strengthen the ground. For example, coarse-grained

granular material can be consolidated to strengths similar to concrete when
cement grout is injected. Indeed, when grouting to enhance strength, a cement-
based grout is preferred if the particle size distribution of the soil permits,
because this gives a higher strength compared with other grouts. If this is not the
case, chemicals must be used since clay-based grouts add little strength to the
soil. Resin- or polymer-based grouts are usually stronger than silicate grouts.
A site investigation should be undertaken prior to planning a grouting
programme. The investigation should provide details of the different soil types
that require grouting, together with their location and thickness. When the object
of grouting is to reduce permeability, porosity and permeability data should be
obtained, as should those relating to hydraulic gradients and the chemistry of the
groundwater. The latter may influence the type of grout chosen and the extent of
treatment, while porosity affects the amount of grout used. The coefficient of
permeability may set the lower limits for permeation grouting. Where ground
strengthening is necessary, additional tests are required such as penetrometer or
pressuremeter tests or laboratory tests on undisturbed samples to assess their
strength.
The determination of the most suitable grout to be injected under given
conditions requires data relating, firstly, to the size of the voids requiring
grouting and, secondly, to the permeability of the ground. The arrangement of
the soil strata to be grouted and the thickness of the overburden should also be
known since these influence the maximum grout pressure that can be used. As
far as the grout itself is concerned, it must be easy to mix and inject—that is,
capable of flowing readily through the void space under reasonable pressure. It
must form a substance with adequate strength when set so that it is not displaced,
GROUTS AND GROUTING 199
and it must set with a minimum amount of shrinkage and remain stable. The
placement of grout should not contaminate the groundwater.
9.2
PROPERTIES OF GROUTS
In order that a grout may bring about the desired effect it is necessary thattttt it
should have the correct fluid properties for injection into the formation, that its
set properties satisfy the design specifications, and that the transformation
between the fluid and final set states should be sufficiently rapid for displacement
of the grout to be unlikely under the stresses to which it will probably be
subjected. An additional set property in the case of hydrogels is volume stability
against syneresis. Other factors which are important in particular circumstances
include the chemical stability of the grout components and their deterioration
under exceptionally high or low site temperatures.
A full description of a grout being evaluated for treatment of a particular soil
includes its initial density, fluid viscosity and shear strength, particle size
distribution of any solids, and any changes in these that may occur during
injection. The possibility of the grout mixing with or being diluted by the
groundwater, as well as the effect of the fluid on the chemistry of the soil and
groundwater, also need to be known.
Viscosity and rigidity are the two rheological properties which govern the flow
of grout in the voids of the soil. In suspension grouts they are inversely
proportional to the water/solids ratio (Figure 9.2(a)). Any increase in the
viscosity of cement grouts with time is sufficiently slow for it to have no effect
on injection. The viscosity of chemical grouts varies with the concentration of
the reactive chemicals (Figure 9.2(b)) and as some of these grouts contain minute
particles in suspension it is perhaps more correct to use the term apparent
viscosity in such cases. Other factors apart (e.g. dilution by groundwater or
reaction with components it carries), gel time depends on the concentration of
activator, inhibitor and catalyst in the grout formulation. With most chemical
grouts, the gel time can be changed by varying the concentration of one or more
of these three components. Increases in temperature can reduce initial viscosity
but the reductions are marginal and are quickly compensated by the accelerated
gelling process.
Grout which has been injected successfully must develop sufficient rigidity to
remain in place. In particular, it must be capable of resisting the hydraulic
pressure exerted by groundwater that has been sealed off by the grouting
operation. Rigidity of an ordinary clay grout cannot be improved by adding more
of the basic clay constituents as viscosity would increase. This is usually done by
adding bentonite (1–10 g/litre) or by using sodium silicate with a reagent.
Sodium silicate tends to thicken the grout and increases rigidity markedly.
The yield point of a grout is the lowest stress that must be applied to initiate
laminar flow. Consequently, it governs the minimum pumping pressure required
Figure 9.2 (a) Viscosities and shear strengths for cement pastes with varying water/
cement ratios (after Raffle and Greenwood, 1961). (b) Viscosities of various grouts; heavy
lines indicate the solution concentrations normally used in the field (after Karol, 1982,
with permission of ASCE)
to inject grout. The yield point of many suspensions increases if they remain at
rest and diminishes when they are agitated. Such suspensions are described as
being thixotropic. Thixotropic behaviour is most notable in the case of bentonite
grouts. One of the problems with thixotropic clay suspensions is that there is no
control over the rate of gelation. Hence the fluid properties of such grouts
generally are modified and their shear strength reduced during injection by the
addition of a dispersing chemical. In this way injection pressures are not
dissipated by the need to overcome yield values but only in overcoming
viscosity. This enhances the rate of flow.
The important properties of fresh particulate grouts are consistency, fluidity,
water retentivity and bleeding. Consistency is a function of resistance to shearing
force and is related to the viscosity coefficient. The fluidity is inversely
proportional to grout density and determines the velocity of flow under given
conditions and given pressure. The more fluid a grout is, the less the pressure
loss when moving so that the grout will travel further from the point of injection.
Generally fluidity is reduced or enhanced according to the amount of water
present. In particulate grouts the resistance to flow is related to the specific
surface of the suspended materials or their shape. High specific surfaces, in the
case of cementitious grouts, impart greater chemical reactivity. Particle shape
can affect flow resistance since spherical shapes tend to reduce friction by
imparting a lubricating type of action. There are two other factors that influence
flow: namely, interparticle attraction or flocculation, which increases flow
resistance, and chemical reactivity of the suspended material in water which, by
gelling or viscosity alteration, also increases flow resistance.
The flow properties of a cement grout are affected primarily by dynamic
interparticle forces of attraction and repulsion, and in dense grouts by dilatancy of
the moving particles. A dense grout can only be pumped easily when it contains
sufficient fluid to prevent expansion of the particle matrix during shear.
Generally a well-graded range of particles is preferred since the better the
grading, the lower the critical porosity at which the grout becomes pumpable. A
reasonable percentage of fine particles is also desirable to increase the specific
surface of the grout particles and thereby slow the separation of solid and liquid
phases.
The movement of fluids through the voids in a soil is resisted by drag at the
interface between the grains and the fluid. For true (Newtonian) fluids the drag is
proportional to viscosity and shear rate (the shear rate being determined from
mean flow velocity and the geometrical characteristics of the void space).
Newtonian fluids possess no shear strength. When a grout which possesses shear
strength is pumped into the ground under a constant pressure, the opposing drag,
due to corresponding shear stress acting at the growing area of the surface wetted
by the grout, eventually equals the applied pressure so that none is available to
maintain flow. In such instances the initial rate of penetration is determined by
the velocity of the grout.
Water retentivity of a grout is indicated by its ability to retain water against
vacuum filtration. In cement grouts, the lower the water/cement ratio, the greater
is the retention of water. The use of water-reducing agents improves water

retentivity.
A particulate grout must be stable, that is, it must be able to remain
homogeneous for a predetermined amount of time. Lack of stability leads to
separation of the constituents, the particles compacting to form a filter and so
giving difficulties during injection. Stability, therefore, is a property of
suspensions, other than thixotropic suspensions, which enables them to preserve
their original rheological characteristics. It depends on the size and specific
surface of the particles along with the intensity of surface action between the
particles and the liquid. In the case of cement grout, the addition of a very small
percentage of bentonite considerably improves its stability. Stability may be
improved by the use of dispersing agents, which also affect fluidity. As far as
clay grouts are concerned, their stability also depends on their thixotropy and
rigidity. Electrolytes, such as potassium nitrate, potassium carbonate, sodium
aluminate, sodium silicate and sodium hydroxide, are commonly used to stabilize
(peptize) clay grouts. Bentonite, in very small concentrations, is also used.
In the case of cement grouts, during mixing the particles are dispersed and
suspended in water. Except in the case of a very dense paste, the resultant
suspension is initially unstable and the cement particles settle under gravity.
Bleeding refers to particles coming out of suspension before the grout has cured.
The bleeding mechanism is important in grout design, particularly the bleeding
rate and the bleeding capacity (i.e. the final volume of bleeding; Figure 9.3). The
greater the degree of dispersion, the smaller is the volume of segregation. If
bleeding water accumulates under individual particles of a cement grout it
destroys their ability to bond and so reduces both the strength and
impermeability of the grouted mass (King and Bush, 1958). Hence low bleeding,
together with active expansion while setting, are important properties.
Expansion of a cement grout, can be brought about by the formation within it
of minute bubbles of hydrogen. These are generated by the addition of aluminium
powder which reacts with the alkalis in the cement. The amount of aluminium
powder required is around 0.01% by weight of the cement (i.e. a few grams).
Expansion improves the bond between the grout and the interstices into which it
is injected.
Cement grouts have well-defined shear strengths that develop immediately
after mixing. The shear strength affects penetrability and Table 9.1 gives values
of limiting penetration for soil permeability of 10−3, 10−4 and 10−5 m/s. The shear
strength, however, is not a limiting factor inhibiting penetration in coarse-grained
soils such as very open gravel.
Figure 9.3 Rates of bleed for ordinary Portland cement (after Littlejohn, 1982, with
permission of ASCE)
Table 9.1 Limiting penetration of shear strengths of cement grouts (after Raffle and
Greenwood, 1961)
Shear strength Limiting Corresponding
(dyne/cm2) penetration for 30 water/cement ratio
m of injection head for ordinary
(m) Portland cement
k=10−3 m/s k=l0−4 m/s k=10−5 m/s
6.6 4.3 1.4 0.5 0.4
25.6 3.6 1.2 0.5
67.6 4.4 0.66
Clay grouts also exhibit shear strength (Figure 9.4) but violent shearing action
temporarily destroys the bulk shear strength. The extent to which it subsequently
redevelops is governed by both the level of subsequent rate of shear and the time
for which the subsequent rate of shear has been maintained.
Cement grouts do not have a controllable gelling time, whereas true fluid
grouts normally do. In some suspensions the gelling action is speeded up by
agitation. This is termed rheopexy and can cause serious trouble in mixing and
injection. Adequate use of dispersants can remedy minor occurrences, but
materials exhibiting a high degree of rheopexy should be abandoned.
Figure 9.4 Development of shear strength in 5% aqueous solutions of two clays after
mixing (after Raffle and Greenwood, 1961)
A clay grout develops a gel after a given interval of time. Nonetheless, even
bentonite grout does not develop a stable gel without the assistance of additivies
such as cement or silicates. The gel is often thixotropic.
The gel time or induction period of a chemical grout is defined as the time
between the initial mixing of the chemical components and the formation of a
gel. The choice of induction period for a specific application depends upon many
factors. Often the two controlling factors are the volume of the grout to be
injected and the pumping rate at the allowable pumping pressure. The
permeability of the ground, groundwater conditions and temperature are also
important factors. The gel time of most chemical grouts can be varied from
minutes to hours but is temperature sensitive. In normal circumstances gel times
of 45–90 min are used to give adequate time for mixing, pumping and
placement. Where setting times are less than 30 min (or ambient temperatures
are high, i.e. greater than 30°C, as setting time is temperature sensitive),
proportioning pump systems may be used which delay mixing of chemical
components until the point of injection. Gel times for very coarse soils should be
relatively short, possibly not exceeding 15 or 30 minutes. On the other hand, in
fine sands and silts, gel times of several hours can be used without seriously
affecting the efficiency of stabilization.
Some chemical grouts, after catalysation, maintain a constant viscosity and at
the end of the induction period turn from liquid to gel almost instantaneously.
Others, such as silicate grouts, maintain their initial viscosity for less than the
induction period and yet others increase in viscosity from the time of catalysis to
gel formation. In the case of these latter two groups of grouts it is not possible to
inject the grout into the soil during the whole of the induction period.
Most applications of grout are into moist ground or below the water table so
that the grouted mass never dries out. Hence the most significant strength factor
is the ‘wet’ strength. However, dry grouted soils possess higher strength than
saturated soils, often by a factor of 10. Apart from the type of grout used, the
strength of a fully grouted soil depends on its density, average grain size and
grain size distribution in that strength increases with increasing density and
decreasing effective grain size (D10). Well-graded soils have higher strengths
when grouted than uniform soils with the same effective grain size. The strength
also depends upon the extent to which the pores in the soil have been filled.
From Table 9.2 it can be seen that the minimum shear strength for a grout in
soil of a given permeability is related to the applied hydraulic gradient. The
demand for a high shear strength in a set grout is greatest in soils of high
permeability. When the hydraulic gradient is very high the choice of grout must
take into account the creep properties of the set grout.
The development of the shear strength of a cement grout is affected by the
water/cement ratio (Figures 9.2(a) and 9.5) and the particle size—the larger the
water content and particle size, the lower the strength. The rate of shear strength
development can be changed by the addition of accelerators or developers.
Under optimum curing conditions, and when not subjected to chemical attack, a
set grout continues to increase in strength over a prolonged period. Cements with
a low rate of hardening tend to have a higher ultimate strength due to slow
formation of denser gel during the initial stages of setting. Generally, ordinary
Portland cement grout has a set strength at 28 days of approximately 60–70% of
ultimate strength (Figure 9.5). This amount of strength may be obtained in about
three days with high early strength cement.
Table 9.2 Hydraulic gradient to maintain flow in non-Newtonian grouts (from Bell, 1975)
Soil permeability (m/s) Yield value (N/m2) Hydraulic gradient
10−2 1 1.2
10 12
100 120
1000 1200
10−3 1 4
10 40
100 400
1000 −
10−4 1 12
10 120
100 1200
1000 −
Soil permeability (m/s) Yield value (N/m2) Hydraulic gradient

10−5 1 40
10 400
100 4000
1000 −
Chemical grouting of granular soils increases their cohesion but the angle of
internal friction remains more or less the same (Figure 9.6). Nevertheless, most
chemical grouts when set form weak solids. On the other hand, some special-
purpose grouts which contain high concentrations of polymerizing components
form very rigid products when set, although at the expense of fluidity during
injection. For instance, crushing strengths typical of strong concretes can be
produced by injection of sands with grouts based on epoxides or polyesters. If
weak grouts are injected into large pores, and a high hydraulic gradient exists,
then they may be extruded bodily.
Chemically grouted soils are likely to be subject to creep. According to Karol
(1982) it is possible to define a creep endurance limit below which failure does
not occur, regardless of the duration of the load. This limit is approximately 25%
of the unconfined compressive strength of the materials concerned in situations
where faces are exposed, as in open excavations, tunnels and shafts. Where
lateral support is afforded to the grouted soil the creep endurance limit
approaches half the unconfined compressive strength.
Shrinkage of cement grout is related mainly to the amount of water removed.
However, shrinkage is not normally a serious problem since the environment into
which grout is injected is often damp or beneath the water table, and moist-cured
grout which remains moist does not shrink, in fact it may expand slightly with
time.
Cement grouts are durable under most normal conditions but deterioration
may be caused by abnormal environmental conditions, such as an attack by
sulphates contained in groundwater or prolonged exposure to salt water.
Durability may be increased by use of blast furnace or sulphate-resisting cement.
Many single-fluid chemical grouts are subject to syneresis, that is, the
expulsion of water from the gel. If this occurs to a significant extent, then
seepage channels develop within the grout itself. In silicate grouts in which
syneresis occurs, water exudes from the gel within a few hours of setting and the
process stabilizes within 3–4 weeks. Syneresis can be controlled in the mix by
increasing the concentration of silicate and the coefficient of neutralization. In
practice the amount of syneresis associated with a given grout depends upon the
ratio of its volume to surface area of the grouted soil since bonding of the gel and
solid surface resists internal shrinkage stresses and so reduces the change in
grout volume. Hence, grouts that may be unsuitable for coarse gravels may be
appropriate for fine sands.
Figure 9.5 (a) Effect of water/cement ratio on strength development of ordinary Portland
cement (OPC) and rapid hardening Portland cement (RHPC). (b) Gain in strength of set
grouts. (From Littlejohn, 1982, with permission of ASCE)
Figure 9.6 Drained triaxial test results for silicate grouted coarse and medium sands (after
Skipp and Renner, 1963)
All grouts that contain water not chemically bound to the grout particles are
subject to mechanical deterioration if subjected to alternative freeze-thaw and/or
wet-dry cycles. The rate at which deterioration occurs varies with the amount of
free water available in the grout, as well as the degree of drying or freezing.
Deterioration of chemical grouts can occur if the grout reacts with the soil or
groundwater (Karol, 1982). Chemical grouts based upon aqueous solutions may
dissolve on long-term contact with groundwater, dissolution governing the
permanence of the grout treatment. The soundest defence against dissolution is to
ensure maximum void filling which precludes access of aggressive groundwater
to grout.
In summary, the choice of grout(s) for a particular project depends on its basic
properties. In the case of particulate grouts, they must be stable during mixing
and injection. Ideally the setting time should be controllable to allow the grout to
set in the right place at the right time. The particle size imposes a lower limit to
the size of void it can penetrate. The viscosity of the grout provides a measure of
the degree to which a grout can penetrate fine pores and its fluidity facilitates
pumping. For grouting by permeation, the lowest viscosity grout that yields
adequate strength tends to be selected for use in formations of moderate to low
permeability. The other flow properties and the gel time determine the maximum
value of the radius of injection. The importance of the gel strength depends on
whether the formation requires strengthening or sealing. The gel strength should
be high enough to resist any tendency to creep. The grout when set must resist
chemical attack and erosion, and displacement by groundwater.
9.3
PENETRATION OF GROUTS
The ability of grout to penetrate soil depends on the particle size distribution,
permeability and porosity of the soil; the pressure being used for injection; the
period of injection; and the viscosity of the grout. Because grouts do not
maintain the same viscosity indefinitely, there is a limit to the distance any grout
can penetrate before the decreasing head available at the advancing interface
between the grout and the pore water, and the increasing viscosity of the grout as
it sets, prevent further flow.
The penetration characteristics of cement or clay grouts or admixtures thereof
are very different from those of chemical grouts. When particulate grouts are
injected into porous soil, filtering may occur whereby the larger particles in the
suspension tend to separate out at the entrances to pores. For example, filtercake
formation occurs when cement grout is injected into uniform sands of less than 2
mm grain size.
Hence the ability of particulate grouts to penetrate a formation depends upon
the particle size of the suspended material in that the particles of the grout must
be smaller than the voids that require filling. This has been indicated in terms of
a groutability ratio, which has been defined as
(9.1)
This ratio should exceed 25 if a grout is successfully to penetrate the formation

concerned. Grouting is not possible if the ratio is less than 11. However, Burwell
(1958) suggested that the D85 value of the grout should not be relied upon solely.
He recommended that when the limits of groutability are approached, the
criterion should be re-examined on the basis of
(9.2)
in order to be doubly sure that the grout is suitable. In the latter case grouting is
consistently possible when N values are above 11 but is impossible when they
are less than 5. Alternatively, the limits for particulate grouts may be taken as a
10:1 size factor between the D15 of the grout and the D15 size of the granular soil
concerned.
Soils containing less than 10% fines can usually be permeated with chemical
grouts. If the fines content exceeds 15%, effective chemical grouting may prove
difficult. Permeation grouting is not possible when the fines content is greater
than 20%. Then hydrofracture must be used.
Particle fineness (given in cm2/g) also offers an indication of groutability. For

example, Nonveiller (1989) quoted limits of penetration for ordinary Portland
cement (OPC); high early strength cement (HESC); colloidal fine cement (CFC);
and ultra-fine cement (UFC) into granular soils. These are given in Table 9.3.
Alternatively, permeability may be used to assess groutability. For example,
Littlejohn (1985) suggested that cement grouts cannot be used for treating soils
with permeabilities less than 5×10−3 m/s as cement particles would be filtered
out during injection. As far as clay grouts are concerned, the small sizes of the
particles permits them to penetrate voids in soil with a permeability of 10−5 m/s.
The limits of groutability of some grout mixes are given in Table 9.4. Initial
estimates of the
Table 9.3 Limits of penetration of cement grout into granular soils

Type of cement Specific surface Permeability, k D85 of cement D15 of soil (mm)
of cement (cm2/ (m/s) (mm)
g)
OPC 3.17 2.3×10−3 0.047 0.87
HESC 4.32 1.3×10−3 0.333 0.67
CFC 6.27 3.2×10−4 0.019 0.38
UFC 8.15 3.5×10−5 0.006 0.12
Table 9.4 Limits of groutability of some grout mixes (after Caron et al., 1975)
Types of soils
Characteristics Coarse sands and Medium to fine sands Silty or clayey
gravels sands, silts
Grain diameter d10>0.5 mm 0.02<d10<0.5 mm d10<0.02 mm
Specific surface S<100 cm 100 cm<S<1000 cm S>1000cm
Permeability k>10−3 m/s 10−3>k>10−5 m/s k<10−5 m/s
Type of mix Bingham Colloid solutions Pure solutions
suspensions (gels) (resins)
Consolidation Cement (k>10−2 m/ Hard silica gels: Aminoplastic
grouting s) double shot: phenoplastic
Aerated mix Joosten (for k>10−4
m/s
—
single shot:
Carongel
Glyoxol
Siroc
Impermeability Aerated mix Bentonite gel Acrylamids
grouting Bentonite gel Ligno-chromate Aminoplastic
Clay gel Light carongel Phenoplastic
Clay-cement Soft silicagel
Types of soils
Characteristics Coarse sands and Medium to fine sands Silty or clayey
gravels sands, silts
Vulcanizable oils
Others (Terranier)
groutability of ground have frequently been based on the results of Lefranc tests.
The penetrability of chemical grouts is governed by their viscosity, injection
pressure and period of injection, as well as the permeability of the soil being
grouted (Bodocsi and Bourers, 1991). Chemical grouts with viscosities less than
2×10−3 N s/m2 (2 cP) (e.g. acrylamide-based grouts) can usually be injected
without difficulty into soils with permeabilities as low as 10−6 m/s. At 5×10−3 N
s/m2 (5 cP) grouts (e.g. chrome-lignin grouts) are limited to soils with
permeabilities higher than 10−5 m/s. At 1×10−2 N s/m2 (10 cP) grouts (e.g.
silicate-based formulations) may not penetrate soils below 10−4 m/s. For higher
viscosities, for example, 2×10−2 N s/m2 (20 cP) it may be necessary to restrict the
grout application to more permeable ground or reduce the hole spacing.
In the case of particulate grouts, penetration is also governed by their shear
strength in that their initial shear strength must be overcome before the grout
begins to flow. The critical shear stress needed to initiate flow of cement and
clay grouts falls within the range 1–20 N/ m2. Although this represents an
extremely weak material, in permeation flow the modest shear stresses involved
must be summed over the very considerable surface area of the passages. Once
started, flow is directly proportional to the excess shear stress.
9.4
TYPES OF GROUT
9.4.1
Cement grouts
Cement grouts may consist of water and cement only, or water, cement and other
materials that combine chemically with the cement for special purposes or that
serve as bulking agents. The most widely used form of grout is a mixture of
ordinary Portland cement and water. Its popularity is due to its ready availability,
its reproducible performance, its high yield strengths, and it is far cheaper than
any chemical grout (Anon., 1990(c)). Its disadvantages are slow gain of strength,
and particle size which precludes permeation into soils of permeability less than
10−3 m/s.
The principal factor affecting the properties of cement grouts is the water/
cement ratio, the amount of water governing the rate of bleeding, subsequent
plasticity and ultimate strength of the grout (Littlejohn, 1982). The extent to
which these and fluidity are related to the water/ cement ratio of neat cement
grout is shown in Figure 9.7. Cement grouts tend to undergo excessive bleeding
at water/cement ratios greater than 1.0. They also have low strength, increased
shrinkage and poor durability. Obviously grouts with high ratios are more difficult
to inject but they undergo less segregation and give higher strength than those
with lower ratios.
Cement grouts have low viscosities and no rigidity for water/cement ratios
higher than 0.5 (by weight). The viscosity of the grout increases sharply and it
acquires some rigidity with lower values of water/cement ratio. For a given
cement, its density after setting is fairly constant and is independent of the water/
cement ratio of the grout. The density varies according to the type of cement
used. Nevertheless, the volume of hardened cement grouts that have identical
proportions and are mixed under similar conditions may vary by as much as 50%.
The setting time of cement grouts increases as the water/cement ratio increases.
For example, cement grouts tend to set after about 4–5 hours, but when they have
been greatly diluted their setting time may amount to 10–15 hours. Some
cements never set for water/cement ratios higher than 10.
Water/cement ratios varying from 0.5:1 up to 10:1 have been used (Anon.
1962). For instance, if the soil contains large voids which accept grout readily,
the water/cement ratio may be 0.5:1. The usual water/ cement ratios fall within
the range 0.8:1 to 5:1.
Pure cement grouts are unstable. As a consequence their flow velocity
diminishes rapidly with increasing distance from the grout-hole and particles
settle out at a rate that decreases with the water/cement ratio of the grout. Hence,
as a general rule the initially injected grout should be fairly thin (water/cement ratio
of 10:1 to 15:1) in order to observe the behaviour of the grout-hole and the
ground, and to minimize the chances of prematurely plugging the hole by too
thick a grout. If the grout-take becomes excessive without building pressure, the
hole should be allowed to rest a sufficient time to allow initial set before
continuing.
In a pure cement grout, the particles of cement are held in suspension only by
the agitation of the mixer or the turbulent flow through the pipes which, for this
purpose, must be of small diameter, generally 25 mm (Cambefort, 1977). Since
the diameter of the grout-hole is larger, the particles of cement gradually block it
up, leaving channels in which the flow remains turbulent. But these channels
may become blocked before the ground is properly injected. In such instances
the refusal pressure which is suddenly attained is artificial. In order to resume
injection it is necessary to redrill the grout-hole.
Figure 9.7 Effect of water content on grout properties (after Littlejohn, 1982, with
permission of ASCE)
Table 9.5 Common additives to cement grout used to impart specific properties to the
final product (after Littlejohn, 1982)
Admixture Chemical Optimum dosage (% Remarks
cement wt)
Accelerator Calcium chloride 1–2 Accelerates set and
hardening
Sodium silicate 0.5–3 Accelerates set
Sodium aluminate 0.5–3 Accelerates set
Retarder Calcium ligno- 0.2–0.5 Also increases fluidity
sulphate
Tartaric acid 0.1–0.5
Sugar 0.01–0.5
Fluidizer Calcium ligno- 0.2–0.3
sulphonate
Detergent 0.05 Entrains air
Admixture Chemical Optimum dosage (% Remarks

cement wt)
Air entrainer Vinsol resin 0.1–0.2 Up to 10% of air
entrained
Expander Aluminum powder 0.005–0.02 Up to 15% preset
expansion
Saturated brine 30–60 Up to l%postset
expansion
Antibleed Cellulose ether 0.2–0.3 (for w<0.7) Equivalent to 0.5% of
mixing water
Aluminum sulphate Up to 20% (for w<5) Entrains air
Cement grout can be improved by mechanical or physico-chemical action.

High-speed mixing (1500–3000 rev/min) activates the hydration of cement
particles, gives a better dispersion and brings about a rapid formation of small
crystalline elements of hydrates of different types. Grouts obtained through
energetic mixing are sometimes called activated or colloidal grouts. Basically
these grouts are more stable and more fluid than ordinary grouts, but only for low
water contents.
Other materials may be added to cement grout to control, to a limited degree,
the physical properties of the mix, such as its shrinkage, its strength and its
setting time (Klein and Polivka, 1958). Basically these admixtures (Table 9.5)
are as follows:
1. Accelerators for hastening the setting time.

2. Retarders for delaying the setting time.
3. Lubricants for increasing flowability.
4. Protective colloids for minimizing segregation.
5. Expansion materials for minimizing shrinkage.
6. Water-reducing materials for reducing the water/cement ratio.
Admixtures for grouts include fillers such as sands, clays and pozzolans. The
main purpose of fillers is to reduce the cost of the grout. Certain fillers, however,
offer certain advantages such as reduced bleeding, or improved fluidity or
retardation. Fine sands can be added to cement grouts when large voids have to
be filled.
Grouts with a very wide range of strengths can be produced by using clay in
combination with cement. Bentonite-cement grouts are stable over almost any
water/cement ratio, provided that a suitable bentonite content is used. The
dispersed bentonite in water acts as a suspending agent preventing cement
particles settling out. Usually about 3% of bentonite causes a marked increase in
viscosity on standing and the tendency is to water down the mix to the
consistency that would be expected with a neat cement grout. Addition of
cement to a thixotropic bentonite-water mix gives it a consistency that varies

from soft plastic to hard rigid strong mortar, depending on the water/cement ratio.
Bentonite holds cement in suspension while hydration takes place, so that the
whole grout sets to form a hard cementitious solid. Owing to its fineness
bentonite has a lubricating quality which allows suspensions of low water/
cement ratio to be pumped, so that little or no surplus water is given off, thus
allowing voids to be filled completely in a single operation.
Pozzolans, notably pulverized fly ash (PFA), when combined with ordinary
Portland cement contribute to the cementitious properties of the hardened grout.
Pulverized fly ash has a similar particle size range to ordinary Portland cement.
The addition of such rounded particles to cement improves the flow properties,
thus helping penetration. The PFA/cement ratios in common use vary from 1:4 to
20:1 depending upon the strength and elastic properties required.
Water-soluble compounds are chemical reagents that are employed in
relatively minute quantities from 0.01 to 2% by weight of ordinary Portland
cement. They include surface-active agents (water-reducing, air-entraining and
dispersing agents), retarders and accelerators. Such agents may be employed
either alone or in combination to enhance the properties of fresh and hardened
grouts.
Surface-active agents improve the fluidity of a grout, which means that the
water/cement ratio may be reduced significantly yet the grout maintains the same
consistency. Such additives can facilitate the penetration of grout. Moreover,
because they lower the water content they decrease the likelihood of bleeding
and improve strength and watertightness. Water-reducing agents include calcium,
sodium or ammonium salts of ligno-sulphates. At higher water/cement ratios
surface-active agents may aid dispersion, whereby particles in the grout are kept
separate. Many surface-active reducing agents (e.g. lignins) retard hydration of
cement, which may be desirable in warm weather. Accelerators, which may be
used in cold weather to hasten setting, include calcium chloride and alkali
carbonates and hydroxides.
When colloidal compounds are added to cement grouts they act as thickeners
and, as such, increase the water requirement of the grout at a given consistency.
They reduce bleeding and segregation significantly in thin grouts.
9.4.2
Cement-clay grouts
Cement-clay grouts are most suitable where the main purpose of grouting is to
arrest water movement (Johnson, 1958). When grouting gravels, some sand or
silt may be included in the mix. Because of the size of cement particles such
grouts are used to treat soils with permeabilities in excess of 10−3 m/s, such as
coarse alluvium. Fluidity is not usually important since the grouts are only used
to treat soils sufficiently open to accept the coarser cement particles without
filtration. The clay is present in the grout in the role of a filler. The development
of strength is slow and there is no well-defined setting time. The set of cement-
clay grout can be accelerated by adding either calcium chloride or sodium
silicate and reagent. The efficiency of these additives is limited when used with
cement-clay grouts having high clay contents, and in such cases the latter
accelerator is used. Set grouts have low crushing strengths in relation to those of
neat cement grouts. Actual strengths range from less than 7 kN/m2 to over 7 MN/
m2, increasing with cement content. The ability of clays to form gels in grout
helps stabilize the cement, significantly slowing down its settlement from
suspension and its bleeding from the grout. The stability of cement-clay grouts is
directly proportional to the quality of the clay and its proportion in the grout. A
small amount of clay is usually required to obtain grouts which settle with little
or no water gain. Grouts with no water gain are keenly sought after since they
usually fully occupy the voids during the first grouting operation so that no
redrilling and regrouting is required. Grouts which settle with no water gain
sometimes can be thixotropic, that is, until the cement hardens. Cement-clay
grouts always have some rigidity. Cement-clay grouts have relatively high yield
values (100–500 dyne/cm2; 10–50 N/m2) which require correspondingly high
injection pressures.
9.4.3
Clay grouts
Natural clays are used whenever possible for clay grouts for reasons of economy.
The particle size distribution of a clay used for grouting is important. Normally a
clay with a liquid limit of less than 60% is not considered unless the coarser
fraction can be removed economically.
The relevant rheological properties of a suitable clay for grouting are more or
less determined by the colloidal content. The higher the content of clay minerals
in a deposit of clay, the better it is for grouting purposes. Montmorillonite is the
most active clay mineral and has the capacity to swell significantly. Viscosity,
rigidity and thixotropy can be affected by adding either electrolytes, colloidal
material (bentonite) or wetting and dispersing agents.
Clay suspensions by themselves are usually slightly alkaline and tend to
flocculate on addition of acidic materials. The pH value of the grout should
exceed 7, if not it should be adjusted by the addition of dispersants such as
sodium phoshate (Kravetz, 1958). Clay grouts with added dispersing chemicals
almost invariably incorporate a third ingredient so that shear strength is restored
as chemical reaction proceeds and the grout can develop a permanent set.
Clay grouts do not have a high strength but reduce the permeability of the
ground and offer high resistance to displacement by water gradients. With neat
clay grouts, particularly bentonite grouts, the shear stress (yield value) that has to
be applied during injection to overcome thixotropic gelling and transform the
suspension to a fluid leads to an unwelcome increase in injection pressures. Use
of a dispersing agent reduces the viscosity, yield stress and gel strength, and makes
the grout more easy to inject. Inclusion of a setting agent can control the set time
within certain limits. The setting agent is normally a sodium silicate.
Bentonite is a moisture-absorbing, colloidal clay which, because of its small
particle size (less than 2 µm), has been used in grouting coarse to medium sands
(Ischy and Glossop, 1962). The gel strength of these grouts is not sufficient to
give appreciable increase in strength to the soil. Hence, they generally are used
only for sealing or reducing permeability. However, the injection of bentonite
grout into sands may be complicated by thixotropic effects, and localized
gelation in small pores can cause flow anomalies. Bentonite is easily pumped when
mixed with water and it can absorb three or more times its dry bulk volume.
At rest bentonite suspension undergo thixotropic gelation. If stirred vigorously
in a mixer, the suspension flows like a mobile liquid, the viscosity decreasing as
the rate of stirring increases. When the motion ceases the suspension begins to set
to a gel which stiffens progressively. Gelation is reversible and the sol-gel
transformation may be repeated indefinitely. The rate of gelation, but not the
final set strength, is influenced markedly by temperature. The gel shrinks
irreversibly on dehydration and on freezing. The strength of a thixotropic gel
depends on the bentonite concentration, the setting time and the chemical
composition of the suspending fluid. The shear strength of a thixotropic
suspension varies with time and may be much higher than its yield value.
Flocculation may be offset by the addition of dispersing agents, thereby
providing a means of controlling the yield value. In addition, dispersing agents
(for example, sodium polyphosphate) extend the range of composition over
which controllable setting can be obtained and may act as retarders.
The addition of sodium silicate to bentonite suspensions leads to a reduction in
yield point and thixotropy, when the clay and silicate concentrations in the
suspension are low. At higher silicate concentrations progressive gelation of the
suspension occurs on standing. The gel formed in this way is permanent and non-
thixotropic. The rate of gelation and the final shear strength of the gel are
markedly influenced by the concentration of bentonite and sodium silicate in the
suspension, and by the shear history of the suspension before it begins to set.
Gels of appreciable mechanical strength can be produced, with setting times of 0.
1–5 hours. No shrinkage occurs during the setting process.
9.4.4
Clay-chemical grouts
Clay-chemical grouts contain clay as the major component. The most common
chemicals used in these grouts are sodium silicate with either sodium aluminate
or hydrochloric acid as reagents (King and Bush, 1963). The chrome-lignins are
also satisfactory gel-forming compounds and are available either as crude
sulphite liquor and chromic acid in liquid form or as a preblended powder that
only needs to be mixed with water. Phosphates may be used to facilitate mixing
of bentonite powder to ensure its full dispersal in reasonable time.
Clay-chemical grouts are stable and possess low yield values and viscosities
which enable them to be injected into fine- to medium-grained sands to reduce
permeability. In fact these grouts represent the lower limit of the suspension
grouts that can be injected into sands. Their ability to penetrate the ground
depends upon initial viscosity and size of clay particles. Bentonite is the finest of
the clays, superbentonite possessing even finer particle grading. Sodium
bentonites react best with chemical additives.
9.4.5
Chemical grouts
Chemical grouts were defined by Anon. (1957) as true solutions which contain
no suspended solid particles unless deliberately added for some specific purpose.
Hence, they are able to penetrate fine sands and sandy silts. However, they are
expensive. They can be used as the final seal after a preliminary stage grouting
has been completed with a cheaper grout. Chemical grouts are Newtonian fluids
and, as such, lack shear strength. They are available in a wide range of
viscosities ranging down to very thin liquids.
Information on the size of openings, the coatings on the surface of openings
and the amount of free water or moisture present must be available before the
type of chemical grout can be chosen. For instance, passages as narrow as 0.05
mm have been grouted and some chemical grouts bond poorly to wet or even
moist surfaces. The pot life (time between mixing of components and the start of
set) of different chemical grouts varies widely from 5 min to several hours. This
must be considered when the grout and method of injection are selected for a
specific project. Furthermore, the temperature of the air and ground must be
known, as they affect the pot life and viscosity of the grout and, hence, the ability
to penetrate voids. Wherever feasible, ground water from the site, at the site
temperature, should be used to prepare the stock solutions to eliminate
differences in tank and underground gel times.
Chemical grouts can be used in two main ways (Karol, 1982). One is a ‘two-
shot’ process in which two different chemicals are brought into contact to form
an insoluble precipitation in the voids. In the two-shot process the more viscous
of the two reactants normally is injected first, followed by injection of a second,
less viscous, chemical which mixes and reacts with the first to produce the
preciptate (for example, the Joosten process uses sodium silicate with calcium
chloride as the gelling agent). Two-shot silicate systems have largely given way
to one-shot systems. In this process all the ingredients are premixed prior to
injection and the grout is so designed that reaction to convert the chemical into a
solid or gelatinous mass takes place in the void (Mitchell, 1970). It is believed
that for both types of grouts virtually all the in-place strength is derived by
mechanical bonding between the treated material and the chemicals.
The two-shot process generally yields much higher unconfined compressive
strengths than the one-shot process. For instance, strengths of between 2.8 and 7.
Figure 9.8 The Joosten method of grouting
0 MN/m2 are commonly associated with the Joosten process. The gel undergoes
almost no shrinkage on setting, it resists temperature changes and saline and
sulphate attack, and is non-toxic. On the other hand, the two-shot process suffers
the disadvantage of being slower because sodium silicate solution is pumped into
the soil as the grout pipe is advanced downwards (Figure 9.8). The pipe then is
flushed with water and calcium chloride is pumped in as the pipe is retracted.
Because precipitation occurs upon contact between the two solutions this limits
the penetrability of silicate grout, necessitating very close spacing of grout pipes.
Higher injection pressures are also required to assist easy acceptance of the
grout. Perfect interpenetration of the two shots may not always be obtained.
Generally, one-shot grouts are more versatile than two-shot systems. Their gel
times may be reproducibly controlled from a few minutes to several hours, with a
wide range of viscosity (1 to 10× 10−3 N s/m2; 1–10 cP) and penetrability (down
to 10−6 m/s) can be obtained (Glossop, 1968). The gel strength of one-shot
silicate grouts is low (up to 3.5 MN/m2). However, if organic esters (ethyl
acetate) are mixed with sodium silicate, not only is the gelling time delayed but
such grout brings about an appreciable increase in strength of the treated soil.
The sodium silicates are the most viscous of the chemical grouts ranging from
2 to 5×10−2 N s/m2 (from 20 to 50 cP). On ageing the gel shrinks, becomes
opalescent and cracks. Syneresis of normally diluted single-phase silicate gels
occurs when they are injected into coarse sands (Figure 9.9). However,
Cambefort (1977) reported that when grains are smaller than 1.5–2 mm, the gel
retains its properties (e.g. the larger the specific surface of sand grains, the higher
the strength of the grouted mass; Figure 9.10). He therefore asserted that it is
imperative to grout a sand-gravel formation with a clay-cement grout to occupy
the large interstices prior to treatment with silicate. Grouts containing 30%
silicate are typical for waterproofing applications. Where high strength is
required, silicate concentrations of 40–60% are used.
Epoxy and polyester resin grouts are generally supplied as two separate
components (base resin and catalyst or hardener) which, when combined with
each other, start to react (polymerize) to form a stiff compound. Various
formulations of polyester resins are available so that the properties of the grout
may be varied by changing the components. Similarly, the pot life of the grout
can be adjusted over a wide range, depending on the temperature, from a few
minutes to several hours. Resin grouts have a considerably lower viscosity than
silicate grouts and develop somewhat higher strength. The compressive strength
of polyester grout ranges up to 140 MN/m2 and over, with tensile strength of up
to 60 MN/m2. These grouts are capable of developing a strong bond with clean,
dry surfaces. However, shrinkage, which can be as high as 6%, occurs during
hardening.
Phenoplast resins are polycondensates resulting from the reaction of phenol on
an aldehyde. Resorcinol is in this category of resins to be used as grout and it is
commonly reacted with formaldehyde (Karol, 1982). A catalyst is required to
control the pH, sodium hydroxide normally being used. Setting times vary
greatly with the pH of the solution, being shortest when the pH is slightly above
9. The initial viscosity of resorcinol formaldehyde ranges from 1.5 to 3×10−3 N s/
m2 (from 1.5 to 3 cP) and it remains more or less constant until gelation starts.
The change from liquid to gel is almost instantaneous.
Aminoplasts are grouts in which the major ingredients are urea and
formaldehyde. Unfortunately these materials will only gel under acid conditions,
a distinct disadvantage to their use as grouts. Hence they can only be used when
the pH of the groundwater is less than 7.
A range of grouts is based on lignin and sodium dichromate. A chrome-lignin
gel is formed by using a dichromate salt to catalyse a calcium ligno-sulphonate
Figure 9.9 Variation of syneresis as a function of grain size; 60% silicate-ethyl acetate
gel (after Caron, 1965)
solution. Water dilution is used and acceleration is produced by dosage with
ferric chloride, which is toxic in the liquid state. A strong elastic gel is produced.
Treated soil strengths range from 170 to 700 kN/m2. Chrome-lignin solutions
have viscosities in the 2 to 5×10−3 N s/m2 (2–5 cP) range and gel times can be
controlled with fair precision in the field. The materials, however, are sensitive to
groundwater dilution and acid conditions. In addition, the viscosity of the
solution increases gradually from the instant of catalysis until a gel is finally
formed.
Acrylamide-based materials come closest in terms of performance to meeting
the specifications for an ideal grout. They penetrate more readily, maintain
constant velocity during the induction period, have better gel time control and
adequate strength for most applications (Karol, 1982). They are, however, more
costly than silicate and acrylamide is neurotoxic. Several acrylamide grouts have
Figure 9.10 Compressive strength of sand-silicate gel mortars. They increase linearly
with the square root of the specific surface of sand grains (from Cambefort, 1977)
been withdrawn because of their toxicity (for example, AM-9, Rocagil BT and
Nitto SS). The gel, however, is non-toxic.
9.4.6
Bitumen emulsion grouts
Bitumen emulsions can be made having different viscosities, rate of break, and
bitumen type and content. An emulsion is a dispersion of minute droplets of a
liquid in another liquid in which it is not miscible. A viscous liquid such as
bitumen can be dispersed in a non-viscous liquid, like water, to give a relatively
non-viscous fluid. Emulsions of bitumen and water may contain 30% of
bitumen. Such dispersions can be injected into soils and then broken, that is, the
two components separate. The breakdown of the emulsion in the soil can be
brought about either by the addition of an organic ester or by means of a
synthetic resin. The time required for the emulsion to breakdown after injection
can be regulated, according to Cambefort (1977) by the use of casein as a
stabilizer and a mixture of sodium aluminate and ethyl formate as a coagulator.

Caustic soda can also be used to extend the injection time and to precipitate and
control the reaction in different water conditions. As the bitumen can be
dispersed as very fine droplets (around 1–2 µm diameter) and as the weight and
viscosity hardly differ from that of water, the emulsion can penetrate into
granular soil (that is, materials with dimensions exceeding 0.1 mm may be
treated). Bitumen is a Newtonian liquid which always flows when subjected to a
pressure gradient. This is a drawback as far as long-term treatment is concerned.
9.5
GROUTING
The design of a successful grouting programme requires the selection of a
suitable grout material, and the correct drilling equipment, procedures and grout-
hole patterns. It is essential that the pipes and injection ports are in the correct
place. It is more important to ensure that the full design soil volume is permeated
with grout when the objective is water cut-off than when the objective is to
improve mechanical properties.
The early stages of a grouting programme are somewhat experimental. Several
holes have to be grouted to provide data relating to the pumping pressures that
may be safely used, the depth of hole to grout in one stage, the grout mixes to
use, the extent of surface leaks and how much grout the injection holes are likely
to accept in a given area. As more data are gathered regarding the geological
conditions and their behaviour under treatment, so a more definite programme
can be planned.
The grout pattern includes the layout of the holes, the sequence in which each
hole is placed and grouted, and the vertical thickness and sequence of grouting
the stages for each hole. The layout of holes may follow some geometric pattern
or this may be modified in relation to the ground conditions. Hole spacings of
about 1.3–2.5m are typical.
Grout-holes in soils may be formed by rotary-percussion, percussion, or
augering methods. Generally, cased holes are required in unconsolidated deposits
where the hole extends to 3m or more. Alternatively, a jetted or driven injection
pipe can be used for grouting. The driving technique has the advantage over
jetting in that it results in a tighter contact between pipe and soil, thereby
minimizing the danger of the grout rising to the surface along the outside of the
pipe rather than penetrating the soil.
The use of gel times that are less than the pumping time, termed fast gel times,
has the advantage of grout location control in flowing groundwater and in
stratified soils. It also limits grout-takes in very pervious materials. Pumping
time to gel time ratios of 10 are common.
The rate at which grout can be injected into the ground generally increases
with an increase in the grouting pressure, but this is limited since excessive
pressures cause the ground to fracture and lift. The safe maximum pressure
depends on the weight of overburden, the strength of the ground, the in-situ
stresses, the pore-water pressures and the permissible amount of ground surface
movement, if any. However, there is no simple relationship between these
factors, and so a common rule-of-thumb for safe maximum grouting pressure is
to relate the pressures to the weight of overburden. For example, the pressures
used, as measured at the top of the hole, may start at 70 kN/m2 for the first 3.1m
stage and increase by 70 kN/m2 in each successive 3.1m stage, while not
exceeding 350 kN/m2 for the fifth and lower stages. Values two to three times as
great may be used with fast gel time systems.
9.5.1
Permeation grouting
Permeation grouting at shallow depths may take place at a single stage from a
grout pipe, the grout-hole being sunk to full depth and then grouted upwards.
Alternatively, grouting may proceed while the hole is drilled. The hole is
extended a short distance using a hollow drill rod, it is then withdrawn this
distance and grout is injected from the rod. The hole remains open over the
length exposed as the soil will have been stabilized sufficiently by the migration
of grout from previous injections. The cycle is then repeated. Lastly, grouting is
continued as the rod is finally withdrawn from the hole.
Stage grouting is used when relatively high grouting pressures have to be
employed to achieve satisfactory penetration of grout in deep holes or tighter
sections of holes. In stage grouting the hole is drilled to a given depth and then
grouted. After the grout has set the hole is deepened for the next stage of
grouting when the procedure is repeated (Figure 9.11). Stage grouting allows
increasing grout pressures to be used for increasing depth of grout-hole and
reduces the loss of grout due to leakage at the surface.
The tube-à-manchette consists of a steel tube, between 37.5 and 62.5 mm in
diameter. This is perforated with rings of small holes (about 8 mm diameter) at
intervals of approximately 0.3m. Each ring of holes is enclosed by a tightly
fitting rubber sleeve which acts as a one-way valve (Figure 9.12). A drillhole is
sunk, with the aid of casing, to the full depth to be treated and the tube-à-manchette
placed in it. The casing is withdrawn and grout, termed sleeve grout (clay-
cement or bentonite), is poured into the annular space left behind. Grouting is
then carried out through the tube-à-manchette by lowering into it a small-
diameter injection pipe perforated at its lower end and fitted with two U-packers.
The packers can be centred over any one of the rings of injection holes. When
injection starts the pressure in the grout pipe rises until the grout lifts the rubber
sleeve, rupturing the sleeve grout and escaping through the small holes into the
soil. The rubber sleeves stop any return of the grout into the tube-à-manchette
and the sleeve grout prevents any leakage of grout at the surface.
Use of the tube-à-manchette offers great flexibility since the same hole can be
grouted more than once and different grouts can be used. In this way coarser
Figure 9.11 Stage-and-packer grouting

soils can be treated first and finer ones later. Hence the tube-à-manchette is often
used to grout alluvial soils.
9.5.2
Claquage grouting
Claquage or fracture grouting is frequently used to grout alluvial soils. The aim
is to develop a network of grouted fractures as the fine-grained types, silts and
clays, are not amenable to penetration grouting (Ischy and Glossop, 1962; Samol
and Priebel, 1985). As injection proceeds claquages spread rapidly, forming an
intermeshing network of grout-filled fractures and in this way reduce the
permeability, as well as compacting and improving the mechanical properties of
the soil. Claquages appear when the pressure in the grout-hole exceeds a certain
value depending on the characteristics of the soil and depth of overburden.
Considerable pressure may be required to start the flow, but once the grout
begins to move the pressure can be reduced. In theory this pressure is
proportional to the rate of flow per unit length of drillhole, to the viscosity of the
grout and inversely proportional to the permeability of the ground. Claquages
may form in any direction. According to Cambefort (1977), vertical fractures
tend to develop before horizontal fractures. The grout occupying horizontal
fractures, in particular, may cause the ground to heave. If damaging heave is to
be avoided, ground movement has to be monitored. Claquage grouting can also
be used for reducing the amount of settlement (e.g. in fills, see Arcones et al.,
1985) or for underpinning to restore the levels of buildings and structures (Raabe
and Esters, 1990).
Figure 9.12 Detail of the tube-à-manchette.
9.5.3
Compaction grouting
Compaction grouting uses highly viscous grout (mixtures of cement, soil, clay
and/or PFA, and water) to compress the surrounding soil (Warner and Brown,
1974). The hardened grout forms a bulb or column of strong, relatively
incompressible material. Although compaction grouting can be used in any type
of soil, it is most frequently used in soils finer than medium-grained sands. One
of its advantages is that its maximum effect is obtained in the weakest soil zones.
As the size of the grout mass increases this can cause the ground to heave. Thus
compaction grouting can be used to correct differential settlements and provide
underpinning of structures. It can also be used to strengthen ground adjacent to
open excavation or tunnelling operations. To be effective, compaction grouting
should not be undertaken at depths less than 1–2 m unless there is an overlying
structure to provide confinement.
Grouting may proceed in stages from near the ground surface, downwards to a
firm-bearing stratum, or from depth to the ground surface. According to Stilley
(1982), if underpinning is the objective, then grouting downwards in stages
beneath the building is the more successful. Grouting from the top downwards
results in a greater grout-take per hole, and a more complete densification of the
lower soils because higher pressures can be used after grouting the overlying
soil. Indeed, when a problem soil extends to the ground surface, downward
grouting is more or less mandatory in order to prevent grout escaping at the
surface. Unless the natural moisture content is substantially on the wet side of
optimum, injection of water prior to grouting results in greater grout-takes.
Typical pressures at the point of injection vary from 350 kN/m2 to 1.7 MN/m2
when injection is within 2m of the surface to over 3.5 MN/m2 when grouting
takes place at depths greater than 6m. Pressures above 4.2 MN/m2 at the injection
point are seldom exceeded irrespective of depth, although pressures as high as 7
MN/m2 sometimes are needed to initiate grouting in a tight hole. It is important
that the pressure build-up during injection is not too rapid. This can be controlled
by adjusting the rate of pumping. A slower pumping rate gives a higher grout-
take.
Pore-water pressures in clay soils should be monitored during grouting. In
some instances slowing the pumping rate prevents the development of excess
pore-water pressures. In others, grouting has to be interrupted to allow
dissipation of the excess pore-water pressure. However, in certain cases
drainage, in the form of band drains or sandwicks, must be provided.
The shape of the grout mass depends upon the character of the soil and the
amount of grout-hole open at the time of injection. In uniform soils the shape is
usually spherical or cylindrical, whereas it tends to be irregular in non-uniform
soils. The size of the grout mass is influenced by the density, moisture content
and mechanical properties of the soil, as well as the rate and pressure of injection.
Masses of grout with diameters of 1m or more are not uncommon (Warner,
1982). Grouting to depths in excess of 30m has been used in compaction
grouting.
9.5.4
Jet grouting
Jet grouting offers a means of forming an impermeable barrier (Coomber, 1986),
as well as providing a means for supporting or underpinning structures. It can be
used in all types of soils and poor soils in relatively inaccessible layers at depth
can be replaced. The quantities of grout involved can be predicted with
reasonable accuracy and close control can be exercised over the zones requiring
treatment. In its simplest form jet grouting involves inserting an injection pipe
into the soil to the required depth (Figure 9.13(a)). The soil is then subjected to a
horizontally rotating jet of water. At the same time the soil is mixed with grout
(cement or cement-bentonite) to form plastic soil-cement. The injection pipe is
gradually raised. Replacement jet grouting involves the removal of soil from the
zone to be treated by a high-energy erosive jet of water and air (Shibazaki and
Ohta, 1982). The water and air are jetted under very high pressure from closely
Figure 9.13 Jet grouting: (a) mixing grout with soil; (b) replacement jet grouting using
compressed air and water to remove soil for simultaneous replacement by grout using
column, panel and wing methods
spaced nozzles at the base of a triple fluid phase drill pipe, the operation
proceeding upwards from the base of a borehole. The grout is emplaced
simultaneously. The soil that is removed is brought to the surface by air lift
pressure. In granular soils some of the coarsest particles are not removed and are
incorporated in the grout. This has proved to be one of the most promising means
of deep stabilization of loess soils. Treatment has been completed successfully to

depths in excess of 40m.
There are three methods of replacement jet grouting: namely, column, panel or
wing grouting (Figure 9.13(b)). However, the rows of columns or panels so
formed cannot be regarded as free-standing or cantilever walls since they do not
possess significant tensile strength and bending resistance. In column grouting,
the columns are formed by rotating the jet pipe as it is raised (Coomber, 1986). A
barrier is produced by overlapping the columns. In panel grouting, jetting takes
place from the pipe in a single vertical plane as it is raised. Interconnecting
panels form cut-offs of low permeability. Wing jetting uses two nozzles, without
rotation, to form a fan-shaped grouted mass. The use of pre-bored guideholes,
usually 150 mm in diameter, facilitates discharge of soil, helps to maintain
verticality and provides a visual check on the continuity of adjacent grouted
areas. Grouting can be terminated at a given level, without disrupting the
overlying soil. The pipe is simply withdrawn, restricted grout flow only being
maintained to top up the guidehole.
The grouts used must possess a low viscosity so that pumping can be
maintained at the required rate. In addition, abrasive fillers such as sand must be
avoided. Apart from these two factors the constituents of a grout mix can be
varied to meet the requirement. For example, if low permeability is the principal
criterion, then cement-bentonites (5% bentonite) normally are used.
Stocker and Zwicker (1987) described the formation of high-pressure injection
columns to reduce differential settlement in the foundation for a highrise
building. The clay soil varied in consistency from soft to stiff. Holes were drilled
for the high-pressure injection columns by continuous flight augers to a depth
exceeding 50m. The holes were installed on a grid pattern at 2.3m centres. The
jet grouting pipes were lowered into the self-supporting uncased holes. A cement
grout with a water cement ratio of 1 was used, the grouting pressure being 400
bars (40 MN/m2), the pull-out speed was 0.2 m/min and the speed of rotation
was 20 rev/min.
A major disadvantage in using jet grouting for pretreatment of the crowns of
tunnels prior to excavation is that stress relief and loosening of the ground can
occur since a void is created before the grout is injected. In order to overcome
this, direct displacement injection has been introduced by removing the water
erosion jet. The grout can be injected on its own or with a stream of air (Anon.,
1991). Columns so formed contain more soil than water-jetted grout columns and
therefore possess a lower strength, but there is less destressing around the
columns. Also the friction between them and the soil is higher and the
surrounding soil is more consolidated. Dugnani et al. (1989) described such sub-
horizontal jet grouting (Figure 9.14) to form a protective zone ahead of a tunnel
face, thereby increasing the stability of the ground and reducing ground
movements to a minimum.
A further innovation is reverse grouting in which grouting is carried out as the
pipe enters the ground instead of as it is withdrawn. The technique requires
Figure 9.14 Sub-horizontal jet-grouting execution sequence for tunnels: (a) jet grouting
ahead of the face; (b) excavation; (c) jet grouting at next stage; (d) geometrical scheme of
treatment. Jet-grouted columns are formed ahead of the tunnel face, the overlapping sub-
horizontal columns forming a protective shell around the excavation. The length of
treatment is normally between 10 and 15 m but excavation stops 2 or 3 m before the end
of the treated zone so as to leave sufficient support for roof stability (after Dugnani et al.,
1989)
higher capacity pumps than conventional jet grouting. In reverse grouting the
nozzles are inclined downwards so that the jet has both horizontal and vertical
components. Provided the soil conditions are suitable (a good rule of thumb is
where the blowcount from a standard penetration test is less than 10) it is
possible to form a grout column directly using the vertical component of the
high-energy jet to displace the soil at the nose of the pipe. The diameter of a
column can be increased from 0.6 (conventional jet grouting) to 1.5m when
grout alone is injected and to 7m when grout is injected with a stream of air.
9.5.5
Some other aspects of grouting
Construction of horizontal grouted diaphragms, which are connected to
impermeable vertical cut-offs, are required when excavations extend below the
water table and groundwater lowering techniques, for one reason or another,
cannot be used and an impermeable formation does not exist at a suitable depth
for the vertical cut-off to be keyed into (Tausch, 1985). Such diaphragms should
be 1–2m in thickness. Because of uplift pressure, they must be constructed at a
level significantly below the proposed base of the excavation in order to prevent
blow-out into the excavation. The grout may be emplaced at the required depth
by a tube-à-manchette. Grout-holes are spaced at 1.1–1.5m, depending upon
ground conditions, and are usually set out in a triangular pattern (Figure 9.15).
The grout injected from each hole overlaps with that injected from the adjacent
holes. Thick diaphragms are formed by grouting in more than one layer.
The reduction of percolation of water through the ground can be accomplished
by forming a grout curtain which involves drilling lines of grout-holes in the
ground to create a barrier or cut-off against excessive seepage (Figure 9.16). The
spacing of grout-holes varies from about 1 to 2.5m apart and grouting takes place
either from an open-pipe or a tube-à-manchette. Once the standard of
permeability has been decided for the whole or a section of a grout curtain, it is
achieved by split spacing or closure methods in which primary, secondary,
tertiary, etc., sequences of grouting are carried out until water tests in the grout-
holes approach the required standard. In multiple row curtains the outer rows
should be completed first, thereby allowing the innermost row to effect closure
on the outer rows. The upstream row should be the tightest row, tightness
decreasing downstream. Single-row curtains are usually constructed by drilling
alternative holes first and then completing the treatment by intermediate holes.
Ideally a grout curtain is taken to a depth where the requisite degree of tightness
is available naturally.
Grouting in water-bearing ground is frequently carried out ahead of a tunnel
face in order to reduce the quantity of water entering the tunnel to readily
manageable amounts (Tan and Clough, 1980). In particular, difficulties
frequently arise when tunnels are excavated beneath rivers which contain buried
channels, especially those occupied by sands and gravels. Such grouting has been
referred to as aureole or umbrella grouting and involves drilling grout-holes in
advance of the tunnel heading which fan out to form a series of concentric
grouted cones (Figure 9.17). The tunnel is then excavated through the grouted
zone. The length of individual holes depends upon the type of ground on the one
hand and the quantity and pressure of water on the other. In most cases, 9m
Figure 9.15 Construction of horizontal grout membrane

should prove sufficient. The type of grout used to form the aureole is governed
by the ground conditions.
Figure 9.16 Serre Poncon dam: (a) typical cross-section; (b) distribution of grout holes in
the central part of the grouted cut-off
Figure 9.17 Auber Station, Paris, second phase of grout treatment

10
Soil stabilization
The objectives of mixing additives with soil are to improve volume stability,
strength and stress-strain properties, permeability, and durability. The
development of high strength and stiffness is achieved by reduction of void
space, by bonding particles and aggregates together, by maintenance of flocculent
structures, and by prevention of swelling. The permeability is altered by
modification of pore size and distribution. Good mixing of stabilizers with soil is
the most important factor affecting the quality of results. The two most
commonly used stabilizers are cement and lime.
10.1
CEMENT STABILIZATION
The addition of small amounts of cement, that is, up to 2%, modify the properties
of a soil, while large quantities cause radical changes in these properties. The
amount of cement needed to stabilize soil has been related to the durability
requirement; put another way, a minimum unconfined compressive strength of 2.
8 MN/m2, after curing at a constant temperature (25°C) and moisture content for
seven days, has been widely used. In fact cement contents may range from 3 to
16% by dry weight of soil, depending on the type of soil and properties required
(Table 10.1). Generally as the clay content of a soil increases, so does the
quantity of cement required.
Any type of cement may be used for soil stabilization but ordinary Portland
cement is most widely used. The two principal factors that determine the
suitability of a soil for stabilization with ordinary Portland cement are, firstly,
whether the soil and cement can be mixed satisfactorily and, secondly, whether,
after mixing and compacting, the soil-cement will harden adequately. Rapid-
hardening cement with extra calcium is used in organic soils, and a retarded
cement will tolerate construction delays. Sulphate-resisting cements are rarely
suitable.
Table 10.1 Typical cement requirements for various soil types (after Anon., 1990(d))
Unified soil Typical range of Typical cement Typical cement
classification cement content for moisture- contents for
requirement,* (% density test (ASTM durability tests
by wt) D 558),† (% by wt) (ASTM D 559 and D
506),‡ (% by wt)
GW, GP, GM, SW, 3–5 5 3–5–7
SP, SM
GM, GP, SM, SP 5–8 6 4–6–8
GM, GC, SM, SC 5–9 7 5–7–9
SP 7–11 9 7–9–11
CL, ML 7–12 10 8–10–12
ML, MH, CH 8–13 10 8–10–12
CL, CH 9–15 12 10–12–14
MH, CH 10–16 13 11–13–15
* Does not include organic or poorly reacting soils. Also, additional cement may be
required for severe exposure conditions such as slope protection.
† ASTM D 558 (1992) Standard Test Method for Moisture-Density Relations of Soil-
Cement Mixtures, American Society for Testing Materials, Philadelphia.
‡ ASTM D 559 (1982) Standard Methods for Wetting and Drying Tests of Compacted Soil-
Cement Mixtures, American Society for Testing Materials, Philadelphia. ASTM
D 506 (1982) Standard Methods for Freezing and Thawing Tests of Compacted
Soil-Cement Mixtures, American Society for Testing Materials, Philadelphia.
10.1.1
Types of soil
Any type of soil, with the exception of highly organic soils or some highly
plastic clays, may be stabilized with cement. Although particles larger than 20
mm diameter have been incorporated in soil-cement, a maximum size of 20 mm
is preferable since this allows a good surface finish. At the other extreme, not
more than about 50% of the soil should be finer than 0.18 mm. Granular soils are
preferred since they pulverize and mix more easily than fine-grained soils and so
result in more economical soil-cement as they require less cement. Typically
soils containing between 5 and 35% fines yield the most economical soil-
cement. As the grain size of granular soils is larger than that of cement, the
individual grains are coated with cement paste and bonded at their points of
contact.
The particles in cohesive soils are much smaller than cement grains and,
consequently, it is impossible to coat them with cement. In practice, cohesive
soils are broken into small fragments which are coated with cement and then
compacted. The hydration products formed after short periods of ageing are
largely gelatinous and amorphous which, with time, harden due to gradual
desiccation. With further curing, poorly ordered varieties of hydrated calcium
SOIL STABILIZATION 237
silicate and hydrated calcium aluminate develop. Ultimately the hydrated cement
forms a skeletal structure, the strength of which depends on the size of the
fragments and amount of cement used. A secondary change occurs in clay soils
due to the free lime in the cement, which reacts with the clay particles, making
the soil less cohesive. However, clay balls may form when the plasticity index
exceeds 8%. Where the soil-cement is exposed to weathering, the clay balls
break down, which weakens the soil-cement. In such instances the US Bureau of
Reclamation requires that clay balls larger than 25 mm diameter are removed
(Anon., 1986b).
It is difficult to mix dry cement into heavy clays and high amounts of cement
have to be added to bring about appreciable changes in their properties. Indeed,
clay soils with liquid limits exceeding 45% and plasticity indices above 18% are
not usually subjected to cement stabilization (Croft, 1968). Heavy clays can,
however, be pretreated with 2–3% cement or, more frequently, with hydrated lime.
This reduces the plasticity, thereby rendering the clay more workable. After
curing for 1–3 days the pretreated clay is stabilized with cement.
Furthermore, the suitability of a clay soil for cement stabilization is controlled
by its texture, and chemical and mineralogical composition. Both kaolinite and
well-crystallized illite have little or no effect on the hydration and hardening
process of cement stabilization. By contrast, the expansive clay minerals,
depending upon their relative activities, may have a profound influence on the
hardening of cement (Bell, 1976). For instance, the affinity of montmorillonite
for lime reduces the pH value of the aqueous phase, and because of the
deficiency in lime the cementitious products developed during curing are inferior
to those of non-expansive clays. This means that the strengths developed are
lower, and unless enough cement is added to supply the free lime requirement to
promote hardening, the properties of the clay are not enhanced. Up to 15% of
cement has to be added to montmorillonitic clays to modify them significantly.
The behaviour of weathered minerals such as degraded illites, chlorites and
vermiculites can be similar to that of montmorillonite. Gibbsite, with its high
response to lime, may retard stabilization of certain lateritic soils. In general,
lime will be more suitable than cement for soils containing these components.
Organic matter and excess salt content, especially sulphates, can retard or
prevent hydration of cement in soil-cement mixtures. In fact, soils containing more
than 2% organic material are usually considered unacceptable (Anon., 1990(d))
and soils with pH values of less than 5 are unsuitable for economic stabilization.
Organic matter retards the hydration of cement because it preferentially
absorbs calcium ions. However, the addition of calcium chloride or hydrated
lime can provide a source of calcium and, consequently, may enable some of
these soils to be treated. In addition, Kuno et al. (1989) found that a mixture of
ordinary Portland cement and 10% (by weight) of gypsum when added to
organic soils gave better results in terms of strength development than did cement
alone.
The disintegration of cement (or lime) stabilized soils, due to sulphate attack,
only occurs when the soil has an appreciable clay fraction and when there is an
increase in moisture content above that at the time of compaction. Sherwood
(1957) maintained that there is a risk of deterioration of clay-cement mixtures
when the content of SO3 and SO4 in the soil is 0.2% and 0.5% or more,
respectively, or if the SO3 content in ground water exceeds 300 mg/l. He found
that sulphate-resistant cement was no better than ordinary Portland cement for
stabilizing clay soils containing sulphates.
10.1.2
Mixing and compaction
In order to achieve a uniform material with minimum cement content, good
standards of mixing are required. Two basic methods of mixing are employed. In
the premix method all the soil is obtained from a borrow pit. Then it is batched,
by weight or volume, into a mixer. After mixing, material is transported to the
site, where it may be spread by hand, grader, stone spreader, concrete spreader or
bituminous paver (Figure 10.1) Compaction is usually by roller. Vibratory rollers
are suitable for granular materials, and dead weight rollers for cohesive materials.
Rubber-tyred rollers appear to operate efficiently on a wide range of materials.
In the mix-in-situ method, mobile mixers are employed to mix the materials in
place on site. If the soil on site is used, savings accrue from a reduction in the
volume of earthworks required. However, in some cases it is more economic to
treat imported materials which are spread, compacted and levelled, prior to
cement spreading and mixing. The machines used for mixing vary from
agricultural rotary tillers to large purpose-built units (Figure 10.2). The smaller
machines are limited to processing depths of 150–200 mm, while some of the
larger machines can process depths in excess of 400 mm. Standards of mixing
are improved, especially in cohesive materials, if rotary tillers are used to
pulverize the soil into small fragments prior to adding cement. It is also often
necessary to add water at this stage. On small jobs bagged cement may be spread
manually, but this is not economic on medium to large size jobs where
mechanical spreaders are used. After mixing in the cement and any additional
water needed to reach the optimum moisture content, compaction and grading to
final level is carried out. Finally, the processed layer is covered with a
waterproof membrane, commonly bitumen emulsion, to prevent drying out and
to ensure cement hydration.
10.1.3
Properties of cement-stabilized soils
The properties developed by compacted cement-stabilized soils are governed by
the amount of cement added on the one hand and compaction on the other. With
increasing cement content the strength and bearing capacity increase, as does the
Figure 10.1 Premixed soil-cement being laid as a sub-base for a road. Note the cover of
bitumen in the foreground which acts as a waterproof membrane
durability to wet-dry cycles. The permeability generally decreases but tends to
increase in clayey soils. Granular soils may become more prone to shrinkage,
whereas the swellability of clay soils is reduced (Table 10.2).
The density achieved is largely a function of compactive effort, soil texture
and, in the case of clay soils, the type of clay minerals present, which determine
the soil moisture response. Adequate compaction is essential for successful
stabilization but prolonged delays between mixing and compaction reduce the
maximum density attainable (Figure 10.3). The addition of cement produces
small increases in the compacted densities of both kaolinitic and illitic clay soils,
but not those containing montmorillonite; in fact the latter gives rise to small
reductions in compaction densities.
Soil-cement undergoes shrinkage during drying and soil-cement made with
different types of soils shows different crack patterns. For example, soil-cement
made with clay develops a higher total shrinkage but crack widths are smaller
and more closely spaced than in soil-cement made with sand. The development of
shrinkage cracks can be reduced by keeping the surface of the soil-cement moist
beyond the normal period of curing and placing it at slightly below optimum
moisture content.
The strength of soil-cement tends to increase in a linear manner with
increasing cement content, but in different soils it increases at different rates
Figure 10.2 Autograder producing mix-in-place stabilized soil (courtesy Bomag Ltd)
(Figure 10.4(a)). Increased pulverization increases the strength of soil-cement.
Ideally the degree of pulverization, if the larger stone particles are excluded,
should break down to parent soil so that 80% of the particles are less than 5 mm
and mixing should be uniform. A lengthy period of mixing brings about partial
hydration of the cement with a resultant loss of strength at constant density. If
compaction is delayed the cement begins to hydrate and therefore the soil-
cement begins to harden. As a result the mixture becomes more difficult to
compact. Compaction should be completed within two hours of mixing. The
strength of soil-cement gradually increases as the time taken in curing increases
(Figure 10.4(b)). Also, the higher the temperature, the more rapid is the gain of
strength. Soil-cement will harden in cold weather providing the temperature does
not fall below 0°C. Excessive drying increases strength but tends to crack the
soil-cement. By contrast, strength is reduced by soaking (Figure 10.5). This is
particularly the case with clayey soils. Typical ranges of 7- and 28-day
unconfined compressive strengths for soaked soil-cement specimens are given in
Table 10.3. It is recommended by Anon. (1990(d)) that specimens are soaked
before being tested since soil-cement may become intermittently or permanently
saturated during its service life and it possesses lower strength under saturated
conditions.
Table 10.2 Typical average properties of soil-cement and soil-lime mixtures (after Ingles
and Metcalf, 1972)
* Variations of approximately 50% around the mean may be expected.

† Values shown are at the additive level optimum for the respective soil types.
‡ Results mav be improved by admixture of the lime with gypsum
Figure 10.3 Loss in strength due to delay in compaction for two soils stabilized with 10%
cement; standard compaction (from Ingles and Metcalf, 1972)
Values of Young’s modulus vary between 140 and 20000 MN/m2 depending
on soil type and increase with increasing cement content. Plastic deformation
occurs on cyclic loading, and under such conditions failure may occur at 60–70%
of the ultimate failure strength. When
Figure 10.4 (a) Relationship between cement content and unconfined compressive
strength for soil-cement mixture; (b) effect of curing time on unconfined compressive
strength of some soil-cement mixtures (from Anon., 1990(d))
Figure 10.5 Effect of soaking on strength (after Ingles and Metcalf, 1972)
subjected to constant loading, soil-cement undergoes creep.
The permeability of most soils is reduced by the addition of cement. In
multiple-lift construction, the permeability along horizontal surfaces is generally
greater than along vertical; for instance, Nussbaum and Colley (1971) found that
permeabilities for flow parallel to the plane of compaction of an embankment were
2–20 times greater than in the plane normal to the direction of compaction.
Table 10.3 Ranges of unconfined compressive strengths of soil-cement (after Anon., 1990
(d))
Soaked compressive strength* (MN/m2)
Soil type 7-day 28-day
Sandy and gravelly soils: Unified groups GW, 2·07–4·14 2·76–6·90
GC, GP, GM, SW, SC, SP, SM
Silty soils: Unified groups ML and CL 1·72–3·45 2·07–6·21
Clayey soils: Unified groups MH and CH 1·38–2·76 1·72–4·14
* Specimens moist-cured 7 or 28 days, then soaked in water prior to strength testing.
10.1.4
Applications of soil-cement
The principal use of soil-cement is as a base material underlying pavements. One
of the reasons soil-cement is used as a base is to prevent pumping of fine-grained
subgrade soils into the pavement above. The thickness of the soil-cement base
depends upon subgrade strength, pavement design period, traffic and loading
conditions and thickness of the wearing surface. Frequently, however, soil-
cement bases are around 150–200 mm in thickness.
Soil-cement has been used to afford slope protection to embankment dams,
soil-cement made from sandy soils giving a durable erosion-resistant facing.
Soil-cement has also provided slope protection for canals, river banks, spillways,
highway and railway embankments and coastal cliffs. Where slopes are exposed
to moderate to severe wave action or rapidly flowing water, the soil-cement
generally is placed in horizontal layers 150–225 mm thick and 2–3 m wide
adjacent to the slope, that is, as ‘stairstep slope protection’. In situations where
conditions are less severe, a layer of soil-cement 150–225 mm thick may be
placed parallel to the slope of the face.
In addition to water storage reservoirs, soil-cement has been used to line
waste-water treatment lagoons, sludge-drying beds, ash-settling ponds and
sanitary landfills. The soil-cement linings are commonly 100–150 mm thick.
Tests sponsored by the US Environmental Protection Agency (Anon., 1983)
showed that when soil-cement was exposed to leachate from sanitary landfills,
the soil-cement hardened considerably and became less permeable. Tests carried
out with hazardous wastes indicated that no seepage had occurred through the
soil-cement after 2.5 years of exposure. Nonetheless, waste materials should be
tested to determine their compatibility with soil-cement. These tests did not
include exposure of soil-cement to acid wastes.
Soil-cement has been used as massive fill replacement to provide uniform
support to foundations where inadequate bearing capacity was available. Such
replacement fills have ranged up to 5.5m in thickness (Dupas and Pecker, 1979).
Soil-cement cushions have been widely used in Bulgaria for the support of
structures on collapsible loess soils (Minkov et al., 1980). The soil-cement
cushion is constructed using loess from the building site. About 3–7% cement is
mixed with the loess and is then compacted in layers between 150 and 200 mm
thick. The cushion is usually 1–1.5 m in thickness, but on some occasions has
exceeded 3m thick. Loess-cement cushions also have been used in conjunction
with heavy tamping.
10.1.5
Plastic soil-cement
On occasions, soil and cement have been mixed at water contents higher than the
consistency of wet concrete and then placed without being compacted. Such
material has been termed plastic soil-cement and a strong durable product can be
obtained using most types of soil except very clayey or organic types.
Cohesionless soils usually form a stronger plastic soil-cement than cohesive
soils. Plastic soil-cement made from silty sand mix may require about 30% water
(this is about twice the desirable water content that would have been used if the
material was compacted with a roller).
In order to achieve comparable strengths, plastic soil-cement needs the
addition of more cement than normal cement-stabilized soil, and its density is
lower. Nevertheless, a compressive strength of 3.5–7.0 MN/ m2 can be attained
after 28 days with most soils. The strength continues to increase as the material
ages, so that after a year it frequently has doubled. Plastic soil-cement, when
cured in moist or saturated ground, tends not to shrink even when it contains a
high percentage of clay.
Plastic soil-cement can be pumped into place to form backfill or tremied into
slurry trenches to form impervious walls. It also has been used to provide thin
linings for irrigation canals.
Plastic soil-cement can be mixed in situ to form shallow piles (9–12 m in
depth) in soils that are not excessively difficult to dig. Cement is injected into the
soil by a special drilling and mixing bit as the hole is advanced, as well as when
the bit is raised and rotated. The resulting columns of plastic soil-cement have not
only proved successful as structural piles but overlapping piles have been
constructed to form underground walls. Normally between 35 and 50% of liquid
cement by volume is added to the volume of the completed soil-cement pile. As
a significant quantity of water flows from the soil-cement into neighbouring soil,
the water/cement ratio of the pile is reduced. This is largely responsible for the
high strengths attained by some piles. Also, the curing conditions are ideal since
the soil-cement is surrounded by either moist or saturated soil at a desirable
temperature.
Another method of forming a mixed in-situ soil-cement pile involves drilling
the hole using water flush, then, as the drill stem is lifted, injecting cement grout
horizontally into the hole from jets in the bit (see section 9.5.4 in Chapter 9). The
mixing is achieved by the turbulent action created by jetting. The pile so formed
has a diameter some three to four times that of the original hole.
10.2
LIME STABILIZATION
Lime stabilization refers to the stabilization of soil by the addition of burned
limestone products, either calcium oxide (i.e. quicklime, CaO) or calcium
hydroxide, Ca(OH)2. On the whole, quicklime appears a more effective stabilizer
of soil than hydrated lime. Moreover, when quicklime is added in slurry form it
produces a higher strength than when it is added in powder form. The process is
similar to cement stabilization except that lime stabilization is applicable to much
heavier clayey soils and is less suitable for granular materials. In fact, the
addition of lime has little effect on soils that contain either a small clay content
or none at all. It also has little effect on highly organic soils.
Lime usually reacts with most soils with a plasticity index ranging from 10 to
50%. Those soils with a plasticity index of less than 10% require a pozzolan for
the necessary reaction with lime to take place, fly ash being commonly used.
Other pozzolans used for the enhancement of lime stabilization include blast
furnace slag and expanded shale. Lime is particularly suited to the stabilization
of heavy clays and may be more effective than cement in clayey gravels. Lime
stabilization of heavy clays gives the soil a more friable structure, which is easier
to work and compact, although a lower maximum density is obtained. The
reaction of lime with montmorillonitic clays is quicker than with kaolinitic clays;
in fact, the difference may amount to a few weeks. A silica surface, however,
should not be considered ‘available’ if it is bound to a similar surface by ions
which are not readily exchangeable. Accordingly, illite and chlorite, although
attacked, are much less reactive than montmorillonite.
When lime is used to stabilize clay soil it forms a calcium silicate gel which
coats and binds lumps of clay together and occupies the pores in the soil.
Reaction proceeds only while water is present and able to carry calcium and
hydroxyl ions to the surfaces of the clay minerals (that is, while the pH value is
high). Consequently, reaction ceases on drying and very dry soils do not react
with lime.
The quantity of lime added should ideally be related to the clay mineral
content as the latter is needed for reaction. Ingles and Metcalf (1972) suggested
that the addition of up to 3% of lime would modify silty clays, heavy clays and
very heavy clays, while 3–4% was required for the stabilization of silty clay, and
3–8% was proposed for stabilization of heavy and very heavy clays. They further
suggested that a useful guide is to allow 1% of lime (by weight of dry soil) for
each 10% of clay in the soil. More exact prescriptions can usually be made after
tests at and slightly each side of the guide value.
Like cement stabilization, lime stabilization tends not to be very effective in
organic soils since the organic matter retards hydration. However, Kuno et al.
(1989) showed that when 20% (by weight) of gypsum was added to quicklime or
hydrated lime, then this mixture could be used to stabilize organic soils as long
as they did not possess excessively high natural moisture contents (Figure 10.6).
The amount of water used in lime stabilization is dictated by the requirements
of compaction. However, if quicklime is used, then extra
Figure 10.6 Soil classification zones based on the natural water content and humic acid
content with guidelines for selecting admixtures on the basis of the soil classification
(after Kuno et al., 1989)
Appropriateness of cementitious admixtures
Soil classification Slaked lime Quicklime Slaked lime + NPC NPC + gypsum
gypsum
Zone A • • • • •
Zone B ° ° • • •
Zone C × ° ° • •
Zone D × × × • •
Zone E × × × × °
Zone F × × × × ×
• Short-term strength increase excellent.
° Short-term strength increase not so good, but develops long-term strength.
× Strength does not develop in long term.
water may be necessary in soils with less than 50% moisture content to allow
for the very rapid hydration process. Furthermore, because the lime-soil reaction
involves exsolution, it is inhibited if the water content of the soil is too low.
Hence moist curing is always desirable.
Mixing is important, and if mixing is delayed after the lime has been exposed
to air, then carbonation of the lime will reduce its effectiveness. Therefore, it is
desirable that mixing be effected as soon as possible and certainly within 24
hours of exposure to air.
10.2.1
Properties of lime-stabilized soils
In most cases the effect of lime on the plasticity of clay soils is more or less
instantaneous. In other words, the plasticity is reduced (this is brought about by
an increase in the plastic limit and reduction in the liquid limit of the soil), as is
the potential for volume change (Figure 10.7). For example, tests carried out by
Figure 10.7 (a) Influence of the addition of lime on the plastic limit, liquid limit and
plasticity of clay of high plasticity. (b) Influence of lime on the linear shrinkage of heavy
and silty clay (from Bell, 1988)
the US Bureau of Reclamation indicated that the addition of 4% lime reduced the
plasticity index of a clay from 47 to 12% and increased the shrinkage limit from
7 to 26% (Anon., 1975). In kaolinitic clay soils, however, lime treatment at times
increases the plasticity index.
The addition of lime to clayey soils increases the optimum moisture content
and reduces the maximum dry density for the same compactive effort
(Figure 10.8). The significance of these changes depends upon the amount of lime
added and the amount of clay minerals present. As lime treatment flattens the
compaction curve, a given percentage of the prescribed density can be achieved
over a much wider range of moisture contents so that relaxed moisture control
specifications are possible.
The strength of soil-lime mixtures depends on several factors such as soil type
and the type and amount of lime added. For example, montmorillonitic clays
give lower strengths with dolomitic limes than with high-calcium or semi-
hydraulic limes (Bell, 1988). Kaolinitic clays, on the other hand, yield the
highest strengths when mixed with semi-hydraulic limes and the lowest strengths
are obtained with high-calcium limes.
Figure 10.8 Influence of the addition of lime on the compaction curves of clay soil (from
Bell, 1988)
Soil mixed with low lime content attains a maximum strength in less time than
that to which a higher content of lime has been added. Strength does not increase
linearly with lime content and in fact excessive addition of lime reduces strength
(Figure 10.9(a)). This decrease is because lime itself has neither appreciable
friction nor cohesion. The optimum lime content tends to range from 4.5 to 8%,
the higher values being required for soils with higher clay fractions. Curing time
is another factor influencing strength of lime-stabilized soil (Bell and Couthard,
1990), a steady gain in strength over months is characteristic (Figure 10.9(b)).
Higher temperatures accelerate curing, and this gives rise to higher strengths
(Figure 10.9(c)). The soil-lime reaction is retarded or may cease once
temperatures fall below 4°C. Hence, strength development more or less ceases
with the onset of cold weather, and strength loss because of cyclic freezing and
thawing is cumulative throughout the winter. Residual strength at the end of
the freeze-thaw period consequently must be sufficient to guarantee the integrity

and stability of the soil-lime layer.
Yet another factor influencing the strength of lime-stabilized soils is their
natural moisture content, the strength decreasing with increasing moisture content.
Even three months after stabilization, the shear strengths of soils with high
saturated moisture contents remain low (Figure 10.9(d)). On the other hand, soil-
lime mixtures compacted at moisture contents above optimum, after brief periods
of curing, attain higher strengths than those compacted with moisture contents
less than optimum. This is probably because the lime is more uniformly diffused
and occurs in a more homogeneous curing environment. The strength of soils
treated at or below optimum moisture content usually can be improved by
spraying each lift with water after compaction.
Al-Rawi and Awad (1981) indicated that the permeability of a clayey soil is
increased when treated with lime as the latter causes flocculation of the soil. The
higher the clay fraction, the more the permeability of the soil-lime mixture
increases. They found that those soils compacted on the dry side of optimum
moisture content developed a higher permeability than those compacted wet of
optimum. They also identified particular moulding moisture contents which gave
minimum permeability for soil-lime mixtures irrespective of the amount of lime
added. However, with increasing age the permeability declines, this being
explained by Brandl (1981) as due to the increased production of gelatinous
reaction products which occupy an increasing amount of void space and
eventually harden.
10.2.2
Uses of lime stabilization
The principal use of the addition of lime to soil is for subgrade and sub-base
stabilization and as a construction expedient on wet sites where lime is used to
dry out the soil. As far as lime stabilization for roadways is concerned,
stabilization is brought about by the addition of between 3 and 6% lime (by dry
weight of soil).
Subgrade stabilization involves stabilizing the soil in place or stabilizing
borrow materials which are used for sub-bases. After the soil, which is to be
stabilized, has been brought to grade, the roadway should be scarified to full
depth and width and then partly pulverized. A rooter, grader-scarifier and/or disc
harrow for initial scarification, followed by a rotary mixer for pulverization, is
employed.
Dry hydrated lime can be spread uniformly by a mechanical spreader or from
bags emptied in piles and levelled off by a drag pulled by a tractor. Some idea of
the amount of lime that is required can be obtained from Figure 10.10. After
sprinkling dry lime with water, preliminary mixing is required to distribute the
lime thoroughly throughout the soil to the proper depth and width and to
Figure 10.9 (a) Effect of the addition of lime, in slurry and powder form, on the strength
of silty clay; (b) Influence of curing time on the unconfined compressive strength of clay
soil of high plasticity treated with different amounts of lime; (c) Influence of curing
temperature and time on strength of clay of very high plasticity stabilized with 5%
addition of lime; (d) Influence of addition of 4% lime on the shear strength of clay soil
with different moisture contents (from Bell, 1988)
Figure 10.10 Rate of application of lime (after Anon, 1982)
pulverize the soil to −50 mm. During mixing, water is added to bring the soil
slightly above the optimum moisture content.
Lime slurries of varying concentrations, depending on the percentage of lime
required and the optimum moisture content, are also applied to the soil. A typical
mix consists of 1 tonne of lime to 2500 litres of water, which produces
approximately 31% lime solution. At higher concentrations there is difficulty in
pumping and handling the slurry spray bars. Forty per cent is a maximum
pumpable slurry. Where low lime percentages are required, the blend may be cut
down to 1 tonne of lime to 3000–3500 litres of water. Since lime in the form of
slurry is much less concentrated than dry lime applications, usually two or more
passes are required to provide the specified amount. Although the slurry method
promotes more uniform distribution of lime, it proves disadvantageous on wet
soils. Its use is generally restricted to projects which require smaller amounts of
lime (4% or under) since, at higher percentages, so much water is required to
carry the lime in solution that the soil would exceed optimum moisture content
most of the time.
After initial mixing the lime-treated soil should be lightly compacted to reduce
carbonation and evaporation loss. Initial curing takes 24–48 hours, water being
added to maintain as near optimum mixture conditions as possible. Where
necessary the soil-lime should be remixed so that it is friable enough for the
required degree of pulverization to be achieved.
Soil-lime mixtures should be compacted to high density in order to develop
maximum strength and stability. This necessitates compacting at or near the
optimum moisture content. Granular bases are generally compacted as soon as
possible after mixing. However, delays of up to two days are not detrimental,
especially if the soil is now allowed to dry out. Clay subgrades can be compacted
soon after final mixing although delays of up to four days are not injurious.
When longer delays (two weeks or more) cannot be avoided, it may be necessary
to incorporate a small amount of extra lime into the soil (for example, 0.5%) to
compensate for losses due to carbonation.
Generally, a five- to seven-day curing period is required. Two types of curing
are employed. Firstly, the soil is kept damp by sprinkling, with light rollers being
used to keep the surface knitted together. Secondly, in membrane curing, the
stabilized soil is either sealed with one application of asphalt emulsion within
one day of final rolling or primed with increments of asphalt emulsion applied
several times during the curing period.
Heavy costs can be incurred when construction equipment and transport
become bogged down on site due to heavy rainfall turning clayey ground into
mud. In these circumstances, an economical method of drying out the top layers
of soil is essential. This can be achieved by the use of quicklime in granular form,
it combining with soil moisture to produce hydrated lime. In the process, heat is
generated which helps the soil dry out.
Lime stabilization is used in embankment construction for roads, railways,
earth dams and levees to enhance the shear strength of the soil. In retaining
structures it is used primarily to increase the resistance to water, either external
or internal. For example, lime has been used to stabilize small earth dams
constructed of dispersive soil and so avoid piping failure. Lime has also been
used to stabilize low-angled slopes, a surface layer of soil about 150 mm thick
being mixed in place.
Lime stabilization of clay soils, especially expansive clay soils, can minimize
the amount of shrinkage and swelling they undergo. Hence, such treatment can
be used to reduce the number and size of cracks developed by buildings founded
on suspect clay soils. Lime stabilization may be applied immediately beneath
strip footings for light structures. The treatment can be better applied as a layer
below a raft in order to overcome differential movement.
The lime slurry pressure injection method involves pumping hydrated lime
slurry under pressure into expansive clay soils (Figure 10.11). The method has
been used to improve the bearing capacity and reduce differential settlements, to
Figure 10.11 Lime slurry injection method (coutesy of Woodbine Corporation, Fort
Worth, Texas)
stabilize failed embankment slopes, to minimize subgrade pumping beneath
railways and as remedial treatment to stabilize soils beneath structures (Joshi et al.,
1981). Usually a depth of 2.1m is sufficient to emplace lime slurry below the
critical zone of changes in moisture content, although depths of up to 40m can be
stabilized. As the slurry is pumped under pressure into the soil, it follows the
path of least resistance, moving along fissures and bedding planes and fractures
created by pumping the slurry. Injection pressures usually range between 345
and 1380 kN/m2, depending on soil conditions, injection taking place until refusal.
If the points of injection are spaced at 1.5m centres, the lime slurry forms a
network of horizontal sheets interconnected by vertical veins. After injection, the
network of soil-lime thickens as a result of calcium ion exchange. This means
that there is a gradual improvement in the swell-shrink behaviour of the soil.
Rogers (1991) has described the use of the lime slurry injection method to
stabilize slopes.
Lime columns, 0.5m in diameter, have been constructed in sensitive clay soils,
to depths of 10m, by means of a tool reminiscent of a giant ‘eggbeater’
(Figure 10.12). The mast and the rotary table of the tool are usually mounted on
a front wheel-loader. A container is attached to the loader to store the quicklime.
The tool is screwed into the soil to the required depth, then the rotation is
Figure 10.12 ‘Eggbeater’ forming lime column (courtesy of Linden-Alimak)

reversed and unslaked lime is forced into the soil, by compressed air, through
openings placed just above the blades of the mixing tool. The amount of lime
used approximates to 5–8% of the dry weight of the soil. The tool is extracted at
a rate about one-tenth of its screwed-in rate so that lime is thoroughly mixed with
the soil. Lime columns can be used to support light structures instead of piles
(Holm et al., 1981).
The bearing capacity of a lime column 0.5m in diameter normally varies
between 50 and 500 kN, depending on the type and amount of lime added. The
strength of a lime column and the increase in strength after installation are
affected by the curing conditions, the confining conditions and pore pressure
may also affect column strength (Ahnberg et al., 1989). The columns reduce both
the total and differential settlements. Another application is around structures
supported by piles, where the lime columns can reduce the negative skin friction
and the lateral displacement of the piles due to creep of the surrounding soil.
Interlocking lime columns have been used around excavations to form a less
rigid block of stabilized soil (Kujala et al., 1985; Saitoh et al.,
1985). Furthermore, lime columns bring about drainage of the soil and compare
favourably with sand drains. In fact Broms and Boman (1976) maintained that
the drainage effect of one lime column was equivalent to that of two or three 100
mm wide band drains or three 150 mm diameter sand drains (see Chapter 4).
Lime piles can be installed in saturated clayey soils by means of a special
metal tube with a closed tip (from 250 to 500 mm in diameter), holes being
Figure 10.13 Installation of lime piles
driven to depths of 5–8 m and spread at 1.5–2.5m centres (Tsytovich et al.,

1971). Then the tube is withdrawn from the soil and the hole is filled with lumps
of quicklime (Figure 10.13). Casing is used if the walls of the hole are unstable.
The casing is extracted as the quicklime is placed. Lastly, the quicklime is
packed by tamping. On contact with the groundwater, the quicklime gradually
slakes, which means that the diameter of the pile may expand from 30 to 70% of
its original size. Furthermore, a large amount of heat is released as slaking takes
place, which causes some of the pore water in the surrounding soil to evaporate.
Lime also penetrates the soil, thereby effecting its stabilization.
10.3
OTHER MATERIALS USED FOR SOIL STABILIZATION
Numerous other materials have been used for stabilizing soil. For example,
pulverized fly ash (PFA) can be used by itself to improve the physical properties
of a soil or in conjunction with lime or cement to form a binder. Pulverized fly
ash is a pozzolan, that is, it reacts with CaO and water to form cementitious
material. Because there is a slower gain in strength when PFA is mixed with
lime, and because PFA has a greater sensitivity to low temperatures, cement has
usually been preferred as the mixing agent. From practical experience, it has
been found advantageous to use rapid-hardening cement. The early gain in
strength provides a safeguard against the effects of adverse weather after

compaction. Pulverized fly ash-cement or PFA-lime stabilization is best suited to
sands and gravels with low clay contents. Well-graded gravel, for example, can
be stabilized by the addition of 10% PFA (based on dry weight) and 5% ordinary
Portland cement.
The compressive strength of the stabilized material is determined by the
characteristics of the PFA, the cement, the degree of compaction and the
efficiency of mixing. Generally PFA stabilizes readily with a reasonably low
cement content. With some mix-in-place procedures, however, there is a risk that
weak areas of poorly stabilized material will be formed if the cement content is
too low. A minimum of 7% is advisable under such conditions. The premixed
method, on the other hand, allows more effective control of cement and moisture
contents.
Lime-pulverized fly ash slurry, injected under pressure has been used to
stabilize embankments and levees in the United States (Baez et al., 1992).
Increases in strength of 15–30% have been recorded.
Bituminous material has been used to stabilize granular soils, providing
cohesion and thereby enhancing their strength. In the case of cohesive soils, the
addition of bitumen waterproofs the soil and in this way the loss of strength
associated with increasing moisture content is reduced. The quantity of bitumen
required for successful waterproofing of cohesive soils normally increases with
increasing clay content. Even so, soils with similar clay contents may require
different amounts of bitumen depending on their affinity for water. Like cement
and lime stabilization, stabilization of organic soils with bitumen is not successful.
Bitumen may be applied hot or as cutbacks, or emulsions. Cutbacks, in which
the bitumen is thinned by the addition of a volatile oil (e.g. paraffin or naphtha),
are more often used than hot bitumen as the latter sometimes proves difficult to
mix with soil. As the volatile evaporates from cutback bitumen, the bitumen
itself is deposited in the soil. The higher the proportion of solvent in cutback
bitumen, the lower is its viscosity and the greater is its volatility, giving a shorter
curing time.
In temperate climates cutback bitumen can be used to stabilize sands since
excess water can be removed by compaction. In the wet-sand mix, 4–10% of
cutback bitumen is mixed with the wet sand to which 1–2% of hydrated lime has
been added previously to assist coating. In the United Kingdom it has been found
that once the silt fraction exceeds 5% the mixture becomes difficult to compact.
In warmer climates 10% or more silt material can be tolerated, and in uniformly
graded sands the figure may rise to 25%.
The amount of cutback bitumen, and hence the volatiles content, required for
attaining maximum strength may be less than that required for most efficient
mixing or compaction. This need for different volatile contents at different stages
of construction can be allowed for by affording the stabilized soil a curing period
during which volatiles evaporate. In other words, a period of time may be
allowed between pulverization and mixing on the one hand and compaction on
the other.
Bituminous emulsions are used cold and consist of a fine suspension of
particles of bitumen in water. The bitumen is deposited as the suspension
coagulates or breaks. Bituminous emulsions are generally only suitable for soil
stabilization in climates where rapid drying conditions occur. Nevertheless,
emulsions have been used in the United Kingdom, for example, to treat cohesive
soils. In such instances cement or lime is added to the mix after treatment with
bitumen. This causes the emulsion to break; it also absorbs some of the excess
moisture as a result of hydration, and affords extra strength to the compacted
soil. About 5–7.5% of emulsion and 3–5% cement or lime are added to soils
treated in this way.
Construction methods used for bituminous stabilization are similar to those
used for soil-cement mixtures. However, the optimum moisture content is
usually below that necessary for efficient mixing. Accordingly, except for sands,
it is often necessary to allow a period of time for the mix to dry before
compaction takes place. In the mix-in-place method, bitumen is usually added in
several passes, each layer being partially mixed before the next pass in order to
avoid saturating the surface of the soil.
Some other soil stabilizers have a useful application in certain special
circumstances, especially where temporary solutions are acceptable. Chemical
additives that have been used for soil stabilization include phosphoric acid,
sodium chloride, calcium chloride, sodium hydroxide, synthetic resins such as
urea formaldehyde, and lignin.
The treatment with phosphoric acid is restricted to acid soils and is normally
ineffective in sands and silts. This is because stabilization is brought about as a
result of acid attack on clay minerals, dissolving their aluminium content. The
latter is then precipitated as hydrated aluminium phosphate.
The addition of sodium chloride is effective in most types of soil, although it
is not so useful in saline or highly organic soils. It aids compaction by prompting
uniformity in the mix and reducing the amount of water loss due to evaporation.
Thus higher densities are attained. However, because of its solubility it is easily
leached from the soil by rain. Consequently, frequent applications are required
unless the surface of the soil is sealed. Soils with high liquid limits usually
respond well to the addition of sodium chloride.
The addition of calcium chloride to soil has similar effects to those of sodium
chloride. However, it has an adverse influence on compaction and tends to
increase the permeability of the soil.
Lateritic soils can be stabilized with the addition of sodium hydroxide, which
is applied with the mix water and facilitates compaction. Sodium hydroxide
reacts with aluminium-bearing minerals, notably kaolinite, and a substantial
increase in strength occurs after an initial curing period. By contrast, the addition
of sodium hydroxide to montmorillonitic soils proves detrimental in that it leads
to a decrease in their strength.
The strength of sandy soils stabilized by urea formaldehyde depends on resin

concentrations, soil characteristics (particularly particle size distribution and clay
content), duration of gelling and pH value of the stabilized medium. In
particular, the strength of such stabilized soils increases as the particle size
distribution decreases, that is, as the permeability is reduced. The presence of
montmorillonite tends to retard hardening and therefore lowers the strength
achievable by such stabilization. In fact, the use of urea formaldehyde is
recommended for stabilization of sand that has less than 3% clay fraction and a
pH value less than 7.5. In such treatment the stabilizing liquid is pumped into
holes in the soil, under pressures of up to 300 kN/m2.
Lignin can be added to soil either in the form of sulphite liquor or as a
powder. Unfortunately, however, lignin is soluble in water and consequently this
type of soil treatment is not permanent. If sodium bichromate or potassium
bichromate is added to sulphite liquor, then an insoluble gel results. Chrome-lignin
has been used to stabilize soil, the chromium ion being an effective chelating
agent, strengthening the bond between clay particles and lignin molecules.
However, such treatment of soils is expensive and is therefore used only in
special situations, for example, in stabilizing volcanic or chloritic soils.
10.4
THERMAL STABILIZATION
Clay soils harden on heating, and if heated to a high enough temperature they
remain hard. This is due to the fact that changes occur in the crystalline structure
of clay minerals above about 400°C, notably the loss of the (OH) group. These
changes are irreversible and give rise to other significant changes in physical
properties.
Basically, thermal stabilization consists of driving exhaust gases, from burning
oil, at temperatures around 1000°C, into holes in the ground (Figure 10.14).
Depths of treatment of up to 20m have been obtained. In the open-firing method
two holes are bored in the soil so that they intersect. The combustion nozzle is
placed over one and the combustion gases exit from the other. In the closed-
firing method a single hole closed system is used in which the burner
temperature is controlled by maintaining an excess air pressure. Air and gas
pressure is typically 25–50% above atmospheric. Gas temperature, ranging from
about 800°C for an air/fuel ratio of 3.5 to 2800°C for a ratio of 1 are achievable.
Stabilization proceeds by progressively locating the flame deeper in the
borehole. Heating is brought about by penetration into the pores of the soil of hot
compressed air. Air temperatures must not exceed the fusion temperature of the
soil. The latter method tends to be the more efficient, although it is not applicable
to soils of low permeability. Many holes can be burned at the same time by
linking a main fuel supply to a series of individually adjusted burner heads.
The distribution of temperature with depth in the soil depends on its porosity,
moisture content and excess pore-water pressure, and the temperature of the
Figure 10.14 Scheme for deep thermal treatment of soil (after Litvinov, 1960): 1,
compressor; 2, cold air pipeline; 3, liquid fuel container; 4, pressurized fuel pump; 5, fuel
pump line; 6, filters; 7, nozzle; 8, combustion chamber; 9, borehole; 10, stabilized zone
gases injected. Until the free water enclosed in the pores of the soil is completely
evaporated, the temperature does not exceed 100°C. Soils containing large
proportions of organic colloids and colloidal minerals react at low temperatures.
For instance, at 200–400°C they begin to develop some water resistance as fine
particles begin to aggregate into granules. Further increase in temperatures leads
to caking and ultimately to fusion. Thermal stabilization cannot be applied to
soils that are saturated since the latent heat of evaporation of water makes it too
expensive.
An alternative method of in-situ heating is to use electric heaters. In one
system compressed air is blown through an electric heater at the top of a
borehole. Temperatures from 500 to 1200°C can be achieved. In another, electric
heaters are inserted in the borehole. Soil around the borehole is baked to form a
column. Walls and structural mats can be formed by interlocking columns. Soils
can be treated regardless of water content or gas permeability.
An appreciable reduction occurs in the plasticity index of a clay if heated to
400°C. Also the moisture absorption capacity of clay is appreciably reduced after
it has been heated to 600°C. Initially the permeability of a clay soil increases on
heating up to 600 or 700°C, above which, owing to the onset of fusion, it
decreases slightly. The compressibility of clay soils is reduced by thermal
treatment, as is their swelling capacity.
Thermal treatment has been used to prevent shear failure in clay soils and to
stabilize clay slopes (Beles, 1957). It also has been used to stabilize collapsible
loess by enhancing its strength appreciably (Litvinov, 1960). Burning oil or gas
in a borehole 0.15–0.2 m in diameter stabilizes a column of loess some ten times
as wide, and depths of 8–10 m can be stabilized in about 10 days. Loess is baked
at a temperature somewhere between 300 and 1000°C (Evstatiev, 1988).
Appendix
Summary of soil-treatment methods
Appendix (continued) Appendix (continued) Appendix (continued) Appendix

(continued)
264
APPENDIX 265
266
Appendix (continued) Appendix (continued)
APPENDIX 267
268
APPENDIX 269
270
References
Ahnberg, H., Bengtsson, P.E. and Holm, G. (1989) Prediction of strength of lime
columns. Proc. 12th International Conf. on Soil Mechanics and Foundation
Engineering, Rio de Janeiro, Vol. 2, pp. 1327–30.
Aldrich, H.P. (1965) Precompression for support of shallow foundations. Proc. American
Society Civil Engineers, Journal Soil Mechanics and Foundations Division, 91
(SM2), 5–20.
Al-Rawi, N.M. and Awad, A.A.A. (1981) Permeability of lime stabilized soils. Proc.
American Society of Civil Engineers, Journal Transportation Engineering Division,
107 (TEI), 25–35.
Andrawes, K.Z., McGown, A., Mashour, M.M. and Wilson-Fahmy, R.F. (1980) Tension
resistant inclusions in soils. Proc. American Society Civil Engineers, Journal
Geotechnical Engineering Division, 106 (GT12), 1313–26.
Anon. (1955) Drainage and erosion control. Subsurface drainage for airfields.
Engineering Manual, Military Construction, Chapter 2, United States Army Corps of
Engineers, Washington, D.C.
Anon. (1957) Chemical grouting. Report of the Task Committee on Chemical Grouting.
Proc. American Society of Civil Engineers, Journal Soil Mechanics and Foundation
Division, 83 (SM4), 1–101.
Anon. (1962) Cement grouting. Progress Report of Task Committee on Cement Grouting.
Proc. American Society Civil Engineers, Journal Soil Mechanics and Foundations
Division, 88 (SM2), 49–98.
Anon. (1975) Earth Manual, US Bureau of Reclamation, Washington, D.C.
Anon. (1976) Methods of Tests for Soils for Civil Engineering Purposes, BS 1377, British
Standards Institution, London.
Anon. (1981a) Code of Practice on Earthworks, BS 6031, British Standards Institution,
London.
Anon. (1981b) Code of Practice on Site Investigation, BS 5930, British Standards
Institution, London.
Anon. (1982) Lime Stabilization Construction Manual, 7th edn, National Lime
Association, Washington, D.C.
Anon. (1983) Lining of Waste Impoundment and Disposal Facilities, Office of Solid
Waste and Emergency Resources, Publication No. SW870, US Environmental
Protection Agency, Washington, D.C.
Anon. (1984) A Geotextile Design Guide, Low Bros, Dundee.
Anon. (1985a) First ‘Stent Wall’ installed at Kingston upon Thames. Ground Engineering,
18 (7), 27–31.
Anon. (1985b) Specifications for cast-in-place diaphragm walling. Ground Engineering,
18 (6), 11–19.
272 REFERENCES
Anon. (1986a) Specification for Highway Works, Department of Transport, HMSO,

London.
Anon. (1986b) Design standards No. 13: Embankment dams. Soil-Cement Slope
Protection, US Bureau of Reclamation, Denver, Chapter 17.
Anon. (1988) Diaphragm walling for Sizewell B sets records. Ground Engineering, 21
(3), 19–25.
Anon. (1989a) Dewatering myths. Ground Engineering, 22 (2), 12–15.
Anon. (1989b) The use of geotextiles and geocomposites in the United Kingdom. Proc.
American Society of Civil Engineers, Journal Construction Engineering and
Management Division, 115 (2), 258–69.
Anon. (1990a) Tropical residual soils, Engineering Group Working Party Report. Quarterly
Journal of Engineering Geology, 23, 1–101.
Anon. (1990b) Geotextile groundwork. Ground Engineering, 23 (3), 14–22.
Anon. (1990c) Cement grouts. Ground Engineering, 23 (7), 18–21.
Anon. (1990d) State-of-the-art report on soil-cement. American Concrete Institute
Materials Journal, 87 (4), 395–417.
Anon. (1991) Jet set style. Ground Engineering, 28 (8), 22–3.
Arcones, A., Ruiz de Temiño, R. and Soriano, A. (1985) Strengthening a structural fill by
claquage. Proc. 11th International Conf. on Soil Mechanics and Foundation
Engineering, San Francisco, Vol. 3, pp. 1677–82.
Arenzana, L., Krizek, R.J. and Pepper, S.F. (1989) Injection of dilute microfine cement
suspensions into fine sands. Proc. 12th International Conf. on Soil Mechanics and
Foundation Engineering, Rio de Janiero, Vol. 2, pp. 1331–4.
Auld, F.A. (1985) Freeze wall strength and stability design problems in shaft sinking.
Proc. 4th International Symp. on Ground Freezing, Sapporo, (eds S. Kinosita and
M. Fukuda), Balkema, Rotterdam, pp. 345–9.
Baez, J.I., Borden, R.H. and Henry, J.H. (1992) Rehabilitation of lower Chariton river
levee by lime/flyash slurry injection, in Transportation Research Record,
Transportation Research Board, National Research Council, Washington, D.C.,
pp. 117–25.
Barker, D.H. (1986) Biodegradable geotextiles for erosion control and slope stabilization.
Civil Engineering, 81 (June), 13–15.
Barron, R.A. (1948) Consolidation of fine grained soils by drain wells. Transactions
American Society Civil Engineers, 133, 718–54.
Baumann, V. and Bauer, G.E.A. (1974) The performance of foundations on various soils
stabilized by the vibrocompaction method. Canadian Geotechnical Journal, 11,
509–30.
Beles, A.A. (1957) Le traitment thermique du sol. Proc. 4th International Conf. on Soil
Mechanics and Foundation Engineering, Paris, Vol. 3, pp. 266–9.
Bell, F.G. (1975) Methods of Treatment of Unstable Ground, Newnes-Butterworths,
London.
Bell, F.G. (1976) The influence of the mineral content of clays on their stabilization with
cement. Bulletin Association of Engineering Geologists, 13, 267–78.
Bell, F.G. (1988) Lime stabilization of clay soils: Part 1, Basic principles. Ground
Engineering, 21 (1), 10–15.
Bell, F.G. and Cashman, P.M. (1986) Groundwater control by groundwater lowering, in
Groundwater in Engineering Geology, Engineering Geology Special Publication No.
REFERENCES 273
3 (eds J.C. Cripps, F.G. Bell and M.G. Culshaw), The Geological Society, London,
pp. 471–86.
Bell, F.G. and Mitchell, J.K. (1986) Control of ground water by exclusion, in
Groundwater in Engineering Geology, Engineering Geology Special Publication No.
3 (eds J.C. Cripps, F.G. Bell and M.G. Culshaw), The Geological Society, London,
pp. 429–43.
Bell, F.G. and Coulthard, J.M. (1990) Stabilization of glacial deposits of the
Middlesbrough Area with cementitious material. Proc. 6th International Congress,
International Association of Engineering Geology, Amsterdam (ed. D.G. Price),
Balkema, Rotterdam, Vol. 3, pp. 1797–807.
Bergado, D.T., Asakami, H. and Alfaro, M.C. (1991) Smear effects of vertical drains on
soft Bangkok Clay. Proc. American Society Civil Engineers, Journal Geotechnical
Engineering Division, 117 (10), 1509–30.
Bertram, G.E. (1940) An Experimental Investigation of Protective Filters, Harvard
University Publication No. 267 .
Bevan, D.J. and Johnson, D. (1989) The use of geogrid to enable treatment of peaty soils
by vibrodisplacement. Ground Engineering, 22 (2), 34–5.
Bjerrum, L. (1972) Embankments on soft ground. General report. Proc. American Society
Civil Engineers, Speciality Conf. on Performance of Earth and Earth Supported
Systems, Vol. 2, pp. 1–54.
Bjerrum, L., Moun, J. and Eide, O. (1967) Application of electro-osmosis to a foundation
problem in a Norwegian quick clay. Geotechnique, 17, 214–35.
Bodocsi, A. and Bowers, M.T. (1991) Permeability of acrylate, urethane and silicate
grouted sands with chemicals. Proc. American Society Civil Engineers, Journal
Geotechnical Engineering Division, 117 (8), 1227–44.
Boyes, R.G.H. (1975) The versatility of diaphragm walling. Civil Engineering, 70
(December), 57–9.
Brand, E.W. and Pang, P.L.R. (1991) Durability of geotextiles to outdoor exposure in
Hong Kong. Proc. American Society Civil Engineers, Journal Geotechnical
Engineering Division, 117 (7), 979–1000.
Brandl, H. (1981) Alteration of soil parameters by stabilization with lime. Proc. 10th
International Conf. on Soil Mechanics and foundation Engineering, Stockholm , Vol.
3, pp. 587–94.
Braun, B., Shuster, J.A. and Burnham, E.W. (1979) Ground freezing for support of open
excavations. Proc. 1st Symp. on Ground Freezing, Bochum (ed. H.L. Jessberger),
Elsevier, Amsterdam, pp. 429–53; also Engineering Geology, 13.
Broms, B.B. and Boman, P. (1976) Stabilization of deep cuts with lime columns. Proc.
6th European Conf. on Soil Mechanics and Foundation Engineering, Vienna, Vol. 1,
pp. 207–10.
Broms, B.B. and Boman, P. (1979) Lime columns. A new foundation method. Proc.
American Society Civil Engineers, Journal Geotechnical Engineering Division, 105
(GT4), 539–56.
Broms, B.B. and Shirlaw, J.N. (1987) Stabilization of waste ponds at Changi depot. Proc.
9th South East Asian Geotechnical Conf., Bangkok, pp. 8/13–8/24.
Brons and De Kruijff, H. (1985) The performance of sand compaction piles. Proc. 11th
International Conf. on Soil Mechanics and Foundation Engineering, San Francisco,
Vol. 3, pp. 1638–86.
274 REFERENCES
Brown, R.E. (1977) Vibroflotation compaction of cohesionless soils. Proc. American

Society Civil Engineers, Journal Geotechnical Engineering Division, 103 (GT12),
1437–51.
Bridle, R.J. (1989) Soil nailing-analysis and design. Ground Engineering, 22(6), 52–6.
Bridle, R.J. and Barr, B.I.G. (1990) Soil nailing. Discussion. Ground Engineering, 23 (5),
30–1.
Bruce, D.A. and Jewell, R.A. (1986) Soil nailing: application and practice—Part 1.
Ground Engineering, 19 (8), 10–15.
Bruce, D.A. and Jewell, R.A. (1987) Soil nailing: application and practice—Part 2.
Burwell, E.B. (1958) Cement and clay grouting of foundations. Practice of the Corps of
Engineers. Proc. American Society of Civil Engineers, journal Soil Mechanics and
Foundation Division, 84 (Paper 1551), 1551/1–1551/22.
Cambefort, H. (1967) Injection des Sols, Eyrolles, Paris.
Cambefort, H. (1977) Principles and applications of grouting. Quarterly Journal of
Engineering Geology, 10, 57–96.
Cannon, E.W. (1976) Fabrics in civil engineering. Civil Engineering, 71 (March), 39–42.
Cantoni, R., Collatta, T., Ghionna, V.N. and Moretti, P.C. (1989) A design method for
reticulated micropile structures in sliding slopes. Ground Engineering, 22 (4), 41–6.
Caron, C. (1965) Etude Physico-Chimique des Gels de Silice. Annals de l’lnstitut du
Batiment et des Travaux Public, No. 207–8.
Caron, C., Cattlin, P. and Herbst, T.F. (1975) Injections, in Foundation Engineering
Handbook (eds H.F.Winterkorn and H.F.Fang), Van Nostrand Reinhold, New York,
pp. 337–53.
Carpentier, R., De Wolf, P., Van Damme, L., De Rouck, J. and Bernard, A. (1985)
Compaction blasting in offshore harbour construction. Proc. 11th International
Conf. on Soil Mechanics and Foundation Engineering, San Francisco, Vol. 3,
pp. 1687–92.
Casagrande, A. (1948) Classification and identification of soils. Transactions American
Society Civil Engineers, 113, 901–92.
Casagrande, L. (1952) Electro-osmosis stabilization of soils. Journal Boston Society Civil
Engineers, 39, 51–83.
Cashman, P.M. (1975) Control of groundwater by groundwater lowering, in Methods of
Treatment of Unstable Ground (ed. F.G. Bell), Newnes-Butterworths, London,
pp. 12–25.
Cashman, P.M. and Haws, E.T. (1970) Control of groundwater by groundwater lowering,
in Ground Engineering (ed. J.S.David), Institution of Civil Engineers, London,
pp. 23–32.
Cedergren, H. (1986) Seepage, Drainage and Flow Nets, 2nd edn, Wiley, New York.
Chi, B.P.C. and Yang, Jiann-Shi (1985) Chemical grouting in Taipei Basin. Proc. 11th
International Conf. on Soil Mechanics and Foundation Engineering, San Francisco,
Vol. 3, pp. 1693–6.
Choa, V. (1989) Drains and vacuum preloading pilot test. Proc. 12th International Conf.
on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Vol. 2,
pp. 1347–50.
Chung, Y.T., Chung, S.T. and Wu, W.K. (1987) Improvements in hydraulic sandy fills by
compaction piles . Proc. 9th South East Asian Geotechnical Conf., Bangkok,
pp. 8/57–8/68.
REFERENCES 275
Clayton, C.R., Simons, N.E. and Matthews, M.C. (1982) Site Investigation, Granada,
London.
Clevenger, W.A. (1958) Experiences with loess as a foundation material. Transactions
American Society Civil Engineers, 123, 151–80.
Cooke, T.F. and Rebenfeld, L. (1988) Effect of chemical composition on physical
structure of geotextiles on their durability. Geotextiles and Geomembranes, 7, 7–22.
Coomber, D.B. (1986) Groundwater control by jet grouting, in Groundwater in
Engineering Geology, (eds J.C.Cripps, F.G.Bell and M.G.Culshaw), Engineering
Geology Special Publication No. 3, The Geological Society, London, pp. 445–54.
Croft, J.B. (1968) The problem of predicting the suitability of soils for cementitious
stabilization. Engineering Geology, 2, 397–424.
D’Appolonia, E.D. (1953) Loose Sands—Their Compaction by Vibroflotation. Special
Technical Publication 156, American Society for Testing Materials, Philadelphia,
pp. 138–54.
D’Appolonia, D.J., Whitman, R.V. and D’Appolonia, E.D. (1969) Sand compaction by
vibratory rollers. Proc. American Society Civil Engineers, Journal Soil Mechanics
and Foundations Division, 95, 263–83.
Dastidor, A.G., Gupta, S. and Gosh, T.K. (1969) Application of sandwicks in a housing
project. Proc. 7th International Conf. on Soil Mechanics and Foundation
Engineering, Mexico City, Vol. 2, pp. 89–94.
De Paoli, B., Mascardi, C. and Stella, C. (1989) Construction and quality control of a 100
m deep diaphragm wall. Proc. 12th International Conf. on Soil Mechanics and
Foundation Engineering, Rio de Janeiro, Vol. 3, pp. 1479–82.
Dugnani, G., Guatteri, G., Roberti, P. and Mosiici, P. (1989) Sub-horizontal jet grouting
applied to a large urban twin tunnel in Campinas, Brazil. Proc. 12th International
Conf. on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Vol. 2,
pp. 1351–4.
Dupas, J.M. and Pecker, A. (1977) Static and dynamic properties of soil-cement. Proc.
American Society of Civil Engineers, Journal Geotechnical Engineering Division,
105 (GT3), 413–36.
Elias, V. (1990) Durability/Corrosion of Soil Reinforced Structures. Federal Highway
Administration, Publication No. FHWA-RD-89–186, Department of Transportation,
McLean, Virginia.
Elias, V. and Juran, I. (1991) Soil Nailing for Stabilization of Highway Slopes and
Excavations, Federal Highway Administration, Report No. FHWA-RD-89–198,
Department of Transportation, McLean, Virginia.
Evstatiev, D. (1988) Loess improvement methods. Engineering Geology, 25, 341–66.
Ferworn, D.E. and Weatherby, D.E. (1992) A Contractor’s Experience with Soil Nailing,
Transportation Research Record, National Research Council, Washington, D.C.
Floess, C.H., Lacy, H.S. and Gerken, D.E. (1989) Artificially frozen ground tunnel: a case
history. Proc. 12th International Conf. on Soil Mechanics and Foundation
Fluet, J.E. (1988) Geosynthetics for soil improvement: a general report and keynote
address, in Geosynthetics for Soil Improvement (ed. R.D.Holtz), Geotechnical Special
Publication No. 18, American Society Civil Engineers, New York, pp. 1–21.
Frivik, P.E. and Thorbergsen, E. (1981) Thermal design of artificial soil freezing systems.
Proc. 2nd International Symp. on Ground Freezing, Trondheim (eds P.E. Frivik,
276 REFERENCES
H.Janbu, R.Saetersdal and L.I.Finorud); Engineering Geology, Vol. 28 (Special

Issue), pp. 189–201.
Fukuoka, M. (1986) Fabric retaining wall with multiple anchors. Proc. 3rd International
Conf. on Geotextiles, Vienna, Vol. 2, pp. 435–40.
Fukuoka, M. and Imamura, Y. (1982) Fabric retaining walls. Proc. 2nd International
Conf. on Geotextiles, Las Vegas, Vol. 2, pp. 575–80.
Gallavresi, F. (1985) Ground freezing-application of mixed method (brine and liquid
nitrogen). Proc. 4th International Symp. on Ground freezing, Sapporo (eds
S.Kinosita and M.Fukuda), Balkema, Rotterdam, Vol. 1, pp. 928–39.
Gambin, M.P. (1987) Dynamic consolidation of volcanic ash in Sumatra. Proc. 9th South
East Asian Geotechnical Conf., Bangkok, pp. 8/145–8/158.
Gassler, G. (1992) German Practice of Soil Nailing, Transportation Research Record,
Transportation Research Board, National Research Council, Washington, D.C.
Gassler, G. and Gudehus, G. (1981) Soil nailing: some aspects of a new technique. Proc.
10th International Conf. on Soil Mechanics and Foundation Engineering,
Stockholm, Vol. 3, pp. 665–70.
Gerber, A. and Harmse, H.J. von M. (1987) Proposed procedure for the identification of
dispersive soils by chemical testing. The Civil Engineer in South Africa, 29, 397–9.
Giroud, J.P. (1982) Filter criteria for geotextiles. Proc. 2nd International Conf. on
Geotextiles, Las Vegas, Vol. 1, pp. 103–8.
Giroud, J.P. (1986) From geotextiles to geosynthetics: a revolution in geotechnical
engineering. Proc. 3rd International Conf. on Geotextiles, Vienna, Vol. 1, pp. 1–18.
Glossop, R. (1968) The rise of geotechnology and its influence on engineering practice.
Geotechnique, 18, 107–17.
Gonze, P., Lejeune, M., Thimas, J.F. and Monjoie, A. (1985) Sand ground freezing for the
construction of a subway station in Brussels. Proc. 4th International Symp. on
Ground Freezing, Sapporo (eds S.Kinosita and M. Fukuda), Balkema, Rotterdam,
Vol. 1, pp. 277–83.
Goto, S. and Iguro, M. (1989) The world’s first high strength, super deep slurry wall.
Proc. 12th International Conf. on Soil Mechanics and Foundation Engineering, Rio
de Janeiro, 3, 1487–90.
Greenwood, D.A. (1970) Mechanical improvement of soils below ground surface. Proc.
Symp. on Ground Engineering, Institution of Civil Engineers, Thomas Telford Press,
London, 11–21.
Gudehus, G. and Schwarz, W. (1985) Stabilization of creeping slopes by dowels. Proc.
11th International Conf. on Soil Mechanics and Foundation Engineering, San
Francisco, Vol. 4, pp. 1697–1700.
Hansbo, S. (1978) Dynamic consolidation of soil by a falling weight. Ground
Hansbo, S. (1979) Consolidation of clay by band-shaped prefabricated drains. Ground
Hansbo, S. (1987) Design aspects of vertical drains and lime column installations. Proc. 9th
South East Asian Geotechnical Conf., Bangkok, Balkema, Rotterdam, pp. 8/1–8/12.
Hansmire, W.H., Russell, H.A., Rawnsley, R.P. and Abbott, E.L. (1989) Field
performance of structural slurry wall. Proc. American Society Civil Engineers,
Journal Geotechnical Engineering Division, 115 (2), 141–56.
REFERENCES 277
Harris, J.S. (1989) State-of-the-art: tunnelling using artificial frozen ground. Proc. 5th
International Symp. on Ground Freezing, Nottingham (eds R.H.Jones and
J.T.Holden), Balkema, Rotterdam, Vol. 1, pp. 245–53.
Hilf, J.W. (1975) Compacted fill, in Foundation Engineering Handbook (eds H.F.
Winterkorn and H.Y.Fang), Van Nostrand Reinhold, New York, pp. 304–11.
Hoare, D.J. (1978) Permeable synthetic fabric membranes. II—Factors affecting their
choice and control in geotechnics. Ground Engineering, 11 (8), 25–31.
Hoare, D.J. (1987) Geotextiles, in Ground Engineer’s Reference Book (ed. F.G. Bell),
Butterworths, London, pp. 34/1–34/18.
Hobbs, N.B. (1986) Mire morphology and properties and behaviour of some British and
foreign peats. Quarterly Journal Engineering Geology, 19, 7–80.
Holm, G., Bredenberg, H. and Broms, B. (1981) Lime columns as foundation for light
structures. Proc. 10th International Conf. on Soil Mechanics and Foundation
Engineering, Stockholm, Vol. 3, pp. 687–94.
Honjo, Y. and Veneziano, D. (1989) Improved filter criterion for cohesionless soils. Proc.
(1), 75–96.
Hughes, J.M.O. and Withers, N.J. (1974) Reinforcing soft cohesive soil with stone
columns. Ground Engineering, 7 (3), 42–9.
Hutchinson, M.T., Daw, G.P., Shotton, P.G. and James, A.N. (1975) The properties of
bentonite slurries used in diaphragm walling and their control. Proc. Conf. on
Diaphragm Walls and Anchorages, London, Institution of Civil Engineers,
pp. 33–40.
Ingles, O.G. and Metcalf, J.B. (1972) Soil Stabilization, Butterworths, Sydney.
Ingold, T.S. (1984a) Geotextiles as filters. Ground Engineering, 17 (2), 29–44.
Ingold, T.S. (1984b) Geotextiles as earth reinforcement in the United Kingdom. Ground
Ingold, T.S. (1991) Partial factor design of polymer reinforced soil walls. Ground
Ischy, E. and Glossop, R. (1962) An introduction to alluvial grouting. Proc. Institution of
Civil Engineers, 22, 449–74.
Jessberger, H.L. (1985) The application of ground freezing to soil improvement in
engineering practice, in Recent Developments in Ground Improvement Techniques
(eds A.S.Balasubramanian, S.Chandra, D.T.Bergado, J.S.Younger and F.Prinzl),
Balkema, Rotterdam, pp. 469–82.
Jewell, R.A. and Pedley, M.J. (1990a) Soil nailing design: the role of bending stiffness.
Jewell, R.A. and Pedley, M.J. (1990b) Soil nailing design: the role of bending stiffness.
Discussion. Ground Engineering, 23 (5), 32–3.
Jewell, R.A., Milligan, G.W.E., Sarsby, R.W. and Dubois, D. (1984a) Interaction between
soil and geogrids. Proc. Symp. on Polymer Grid Reinforcement in Civil Engineering,
London, Institution of Civil Engineers, Thomas Telford Press, London, Paper 1.3.
Jewell, R.A., Paine, N. and Woods, R.I. (1984b) Design methods for steep reinforced
embankments . Proc. Symp. on Polymer Grid Reinforcement in Civil Engineering,
London, Institution of Civil Engineers, Thomas Telford Press, London, pp. 70–81.
Johnson, S.J. (1958) Cement and clay grouting of foundations: grouting with clay-cement
grouts. Proc. American Society of Civil Engineers, Journal Soil Mechanics and
Foundations Division, 84 (SM1, Paper 1545), 1545/1–1545/12.
278 REFERENCES
Johnson, S.J. (1970) Precompression for improving foundation soils. Proc. American
Society Civil Engineers, Journal Soil Mechanics and Foundations, 96 (SM1),
145–75.
Jones, J.S. (1981) Engineering practice in artificial ground freezing. State-of-the-art
report. Engineering Geology (Special Issue), 18, 313–26.
Jones, C.J.F.P. (1985) Reinforced Earth, Butterworths, London.
Jones, C.J.F.P., Murray, R.T., Temporal, J. and Mair, R.J. (1985) First application of
anchored earth. Proc. 11th International Conf. Soil Mechanics and Foundation
Joshi, R.C., Natt, G.S. and Wright, P.J. (1981) Soil improvement by lime-fly ash slurry
injection. Proc. 10th International Conf. on Soil Mechanics and Foundation
Engineering, Stockholm, Vol. 3, pp. 707–12.
Juran, I., Ider, H.M. and Farrag, K. (1990a) Strain compatibility analysis for geosynthetic
reinforced soils walls. Proc. American Society of Civil Engineers, Journal
Geotechnical Engineering Division, 116 (2), 303–29.
Juran, I., Baudrand, G., Farrag, K. and Elias, V. (1990b) Design of soil nailed retaining
structures, in Design and Performance of Earth Retaining Structures (eds P.C.Lambe
and L.A.Hansen), American Society Civil Engineers, Geotechnical Special
Publication No. 25, New York, pp. 644–59.
Kamauchi, H. and Kitamori, I. (1985) Improvement of soft ground bearing capacity using
synthetic meshes. Journal of Geotextiles and Geomembranes, 2, 3–32.
Karol, R.H. (1968) Chemical grouting technology. Proc. American Society Civil
Engineers, Journal Soil Mechanics and Foundations Division, 94 (SM1), 175–204.
Karol, R.H. (1982) Chemical grouts and their properties, in Grouting in Geotechnical
Engineering, Speciality Conf., New Orleans (ed. W.H.Baker), American Society
Civil Engineers, New York, pp. 359–77.
King, J.C. and Bush, E.G.W. (1963) Grouting of granular materials. Symp. on Grouting,
Transactions American Society of Civil Engineers, 128, 1279–1310.
Klein, A.M. and Polivka, M. (1958) Cement and clay grouting of foundations: on the use
of admixtures in cement grouts. Proc. American Society of Civil Engineers, Journal
Soil Mechanics and Foundations Divisions, 84 (SM1, Paper 1547), 1547/1–1547/24.
Klein, J. (1989) State-of-the-art: engineering design of shafts. Proc. 5th International Symp.
on Ground Freezing, Nottingham (eds R.H. Jones and J.T. Holden), Balkema,
Rotterdam, Vol. 1l, pp. 235–44.
Koerner, R.M. and Robins, J.C. (1986) In situ stabilization of slopes using nailed
geosynthetics. Proc. 3rd International Conf. on Geotextiles, Vienna, Vol. 2,
pp. 345–400.
Kravetz, G.A. (1958) Cement and clay grouting of foundations: the use of clay in pressure
grouting. Proc. American Society of Civil Engineers, Journal Soil Mechanics and
foundation Division, 84 (SM1, Paper 1546), 1546/1–1546/30.
Kujala, K. (1989) Frost action and the mechanical properties of an artificially frozen test
plot. Proc. 12th International Conf. on Soil Mechanics and Foundation Engineering,
Rio de Janeiro, Vol. 2, pp. 1449–54.
Kujala, K., Halkola, H. and Lahtinen, P. (1985) Design parameters for deep stabilized soil
evaluated from in situ and laboratory tests . Proc. 11th International Conf. on Soil
Mechanics and Foundation Engineering, San Francisco, Vol. 3, pp. 1717–20.
REFERENCES 279
Kuno, G., Kutara, K. and Miki, H. (1989) Chemical stabilization of soft soils containing
humic acid. Proc. 12th International Conf. on Soil Mechanics and Foundation
Lawson, C.R. and Ingles, O.G. (1982) Long term performance of MESL road sections in
Australia. Proc. 2nd International Conf. on Geotextiles, Las Vegas, Vol. 2,
pp. 535–9.
Lee, S.L., Karunaratne, G.P., Yong, K.Y. and Ramaswamy, S.D. (1989) Performance of
fibre drain in consolidation of soft soils. Proc. 12th International Conf. on Soil
Mechanics and Foundation Engineering, Rio de Janeiro, Vol. 3, pp. 1667–70.
Leflaive, E. (1988) Texsol: already more than 50 successful applications. Proc.
International Symp. on Theory and Practice of Earth Reinforcement, Tokyo,
pp. 541–5.
Leshchinsky, D. and Boedeker, R.H. (1989) Geosynthetic reinforced soil structures. Proc.
American Society of Civil Engineers, Journal Geotechnical Engineering Division,
115 (10), 1459–78.
Lewis, W.A. and Parsons, A.W. (1958) An Investigation of the Performance of a 33/4 ton
Vibrating Roller for Compacting Soil, DSIR Road Research Laboratory, Research
Note No. RN/3219/WAL.AWP, Crowthorne.
Lippomann, R. and Gudehus, G. (1989) Dowelled clay slopes: recent examples. Proc.
12th International Conf. on Soil Mechanics and Foundation Engineering, Rio de
Janeiro, Vol. 2, pp. 1269–71.
Littlejohn, G.S. (1970) Soil anchors, in Ground Engineering (ed. J.S. Davis), Thomas
Telford Press, London, pp. 33–45.
Littlejohn, G.S. (1982) Design of cement based grouts, in Grouting in Geotechnical
Engineering, Speciality Conf., New Orleans (ed. W.H. Baker), American Society
Civil Engineers, New York, pp. 35–48.
Littlejohn, G.S. (1985) Chemical grouting. Ground Engineering, 18: Part 1, No. 2,
pp. 13–18; Part 2, No. 3, pp. 23–8; Part 3, No. 4, pp. 29–34.
Littlejohn, G.S. (1990) Ground anchorage practice, in Design and Performance of Earth
Retaining Structures (eds P.D.Lambe and L.A.Hansen), American Society Civil
Engineers, Geotechnical Special Publication No. 25, pp. 692–733.
Litvinov, I.M. (1960) Stabilization of settling and weak clayey soils. Highway Research
Board Special Report No. 60 , Washington, D.C., pp. 94–112.
Litvinov, I.M. (1973) Deep compaction of soils with the aim of considerably increasing
their bearing capacity. Proc. 8th International Conf. on Soil Mechanics and
Foundation Engineering, Moscow, Vol. 3, pp. 392–4.
Lizzi, F. (1977) Practical engineering in structurally complex formations. The in situ
reinforced earth. Proc. International Symp. on Geotechnics in Complex Formations,
Capri, pp. 327–33.
Lo, K.Y., Inculet, I.I. and Ho, K.S. (1991) Electro-osmotic strengthening of soft sensitive
clay. Canadian Geotechnical Journal, 28, 62–73.
Lovell, C.W. (1957) Temperature effects on phase composition and strength of partially
frozen ground. Highway Research Board, Bulletin No. 168, Washington, D.C.
Madu, R.M. (1977) An investigation into the geotechnical properties of some laterites of
eastern Nigeria. Engineering Geology, 11, 101–25.
Manfakh, G.R. (1989) Soil reinforcement—a tale of three walls. Proc. 12th International
Conf. on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Vol. 2,
pp. 1285–8.
280 REFERENCES
Mansur, C.I. and Kaufman, R.F. (1962) Dewatering, in Foundation Engineering (ed.
G.A. Leonards), McGraw-Hill, New York, pp. 241–350.
McGown, A. (1971) The classification for engineering purposes of tills from moraines
and associated landforms. Quarterly Journal of Engineering Geology, 4, 115–30.
McGown, A. (1976) The properties and uses of permeable fabric membranes. Proc.
Residential Workshop on Materials and Methods for Low Cost Road, Rail and
Reclamation Work, Leura, Australia, pp. 663–709.
McGown, A. and Hughes, F.H. (1981) Practical aspects of the design and installation of
deep vertical drains. Geotechnique, 31, 3–18.
McGown, A., Kabir, M.H. and Murray, R.T. (1978) Compressibility and hydraulic
conductivity of geotextiles. Proc. 1st International Conf. on the Use of Fabrics in
Geotechnics, Paris, Vol. 2, pp. 167–72.
McGown, A., Andrawes, K.Z. and Kabir, M.H. (1982) Load extension properties of
geotextiles confined in soil. Proc. 2nd International Conf. on Geotextiles, Las Vegas,
Vol. 3, pp. 793–8.
McGown, A., Andrawes, K.Z., Hytiris, N. and Mercer, F.B. (1985) Soil strengthening
using randomly distributed mesh elements. Proc. 11th International Conf. on Soil
Mechanics and Foundation Engineering, San Francisco, Vol. 3, pp. 1735–8.
McKeand, E. and Sissons, C.R. (1978) Textile reinforcements—characteristic properties
and their measurements. Ground Engineering, 11 (5), 13–16.
McKittrick, D.P. (1979) Reinforced earth: application of theory and research to practice.
Menard, L. and Broise, Y. (1975) Theoretical and practical aspects of dynamic
consolidation. Geotechnique, 15, 3–18.
Mettier, K. (1985) Ground freezing for the construction of the Milchbruch tunnel in
Zurich, Switzerland. Proc. 4th International Symp. on Ground Freezing, Sapporo
(eds S.Kinosita and M.Fukuda), Balkema, Rotterdam, Vol. 1, pp. 263–70.
Meyerhof, G.G. (1951) Building on fills with special reference to settlement of a large
factory. Structural Engineer, 29 (11), 297–305.
Miller, E. (1988) The eductor dewatering system. Ground Engineering, 21 (6), 29–34.
Millet, R.A. and Perez, J.Y. (1981) Current USA practice: slurry wall specifications. Proc.
(GT8), 1041–56.
Minkov, M., Evstatiev, D. and Donchev, P. (1980) Dynamic compaction of loess. Proc.
International Conf. on Compaction, ENPC-LCPC, Paris, pp. 345–9.
Mitchell, J.K. (1970) In place treatment of foundation soils. Proc. American Society Civil
Mitchell, J.K. (1981) Soil improvement—State-of-the-art report. Proc. 10th International
Conf. on Soil Mechanics and Foundation Engineering, Stockholm, Vol. 3,
pp. 509–65.
Mitchell, J.K. and Wan, T.Y. (1977) Electro-osmotic consolidation—its effect on soft
soils. Proc. 9th International Conf. on Soil Mechanics and Foundation Engineering,
Tokyo, Vol. 1, pp. 219–24.
Mitchell, J.K. and Villet, W.C.B. (1987) Reinforcement of Earth Slopes and
Embankments, National Cooperative Highway Research Program, Report 290,
Transportation Research Board, National Research Council, Washington, D.C.
Mitchell, J.K. and Christopher, B.R. (1990) North American practice in reinforced soil
systems, in Design and Performance of Earth Retaining Structure (eds P.C.Lambe
REFERENCES 281
and L.A.Hansen), American Society Civil Engineers, Geotechnical Special

Publication No. 25, New York, pp. 322–46.
Moh, Z.C. and Woo, S.M. (1987) Preconsolidation of Bangkok Clay by nondisplacement
sand drains and surcharge. Proc. 9th South East Asian Geotechnical Conf., Bangkok,
pp. 8/171–8/174.
Moorhouse, D.C. and Baker, G.L. (1969) Sand densification by heavy vibratory
compactor. Proc. American Society Civil Engineers, Journal Soil Mechanics and
Foundations Div., 95, 985–94.
Nonveiller, E. (1989) Grouting Theory and Practice, Elsevier, Amsterdam.
North-Lewis, J.P. and Lyons, G.H.A. (1975) Continguous bored piles. Proc. Conf. on
Diaphragm Walls and Anchorages, London, Institution Civil Engineers, pp. 189–94.
Nussbaum, P.J. and Colley, B.E. (1971) Dam Construction and Facing with Soil-Cement,
Research and Development Bulletin No. RDOIOW, Portland Cement Association,
Skokie, 14p.
O’Rourke, T.D. and Jones, C.J.F.P. (1990) Overview of earth retention systems: 1970–
1990, in Design and Performance of Earth Retaining Structures (eds P.C.Lambe and
L.A.Hansen), American Society Civil Engineers, Geotechnical Special Publication
No. 25, pp. 22–51.
Parsons, A.W. (1987) Shallow compaction, in Ground Engineer’s Reference Book (ed.
F.G.Bell), Butterworths, London, pp. 37/3–37/17.
Pedley, M.J., Jewell, R.A. and Milligan, G.W.E. (1990) A large scale experimental study
of soil-reinforced interaction. Ground Engineering, 20: Part 1, No. 5, pp. 44–50; Part
2, No. 6, pp. 45–9.
Perry, W. (1963) Electro-osmosis dewaters large foundation excavation. Construction
Method, 45 (9), 116–19.
Placzek, D. (1989) Methods for the calculation of settlements due to groundwater
lowering. Proc. 12th International Conf. on Soil Mechanics and Foundation
Engineering, Rio de Janeiro, Vol. 3, pp 1813–18.
Powers, J.P. (1981) Construction Dewatering: A Guide to Theory and Practice, Wiley
Interscience, New York.
Powrie, W. and Roberts, T.O.L. (1990) Field trials of an ejector well dewatering system
at Conwy, North Wales. Quarterly Journal of Engineering Geology, 23, 169–85.
Prugh, B.J. (1960) New tools and techniques for dewatering. Proc. American Society
Civil Engineers, Journal Construction Division, 80 (CO1), 11–25.
Prugh, B.J. (1963) Densification of soils by explosive vibration. Proc. American Society
Civil Engineers. Journal Construction Division, 83 (CO1), 79–100.
Raabe, E.W. and Esters, K. (1990) Soil fracturing techniques for terminating settlements
and restoring levels of buildings and structures. Ground Engineering, 23 (3), 33–45.
Raffle, J.F. and Greenwood, D.A. (1961) The relationship between the rheological
characteristics of grouts and their capacity to permeate soils. Proc. 5th International
Conf. on Soil Mechanics and Foundation Engineering, Paris, Vol. 2, pp. 789–93.
Rankilor, P.R. (1981) Membranes in Ground Engineering, Wiley, Chichester.
Rathmayer, H.G. and Saari, K.H.O. (eds) (1983) Improvement of ground. Proc. 8th
European Conf. on Soil Mechanics and Foundation Engineering, Helsinki, Balkema,
Rotterdam.
Ressi, A. and Cavalli, N. (1984) Bentonite slurry trenches. Engineering Geology, 21,
333–9.
282 REFERENCES
Richart, F.E. (1959) A review of theories for sand drains. Proc. American Society Civil
Engineers, Journal Soil Mechanics and Foundations Division , 83 (SM3), 1301–38.
Robinson, K.E. and Eivemark, M.M. (1985) Soil improvement using wick drains and
preloading. Proc. 11th International Conf. on Soil Mechanics and Foundation
Rogers, C.D.F. (1991) Slope stabilization using lime. Proc. International Conf. on Slope
Stability Engineering, Developments and Applications, Isle of Wight, Thomas
Telford Press, London, pp. 395–402.
Rowe, P.W. (1964) The calculation of the consolidation rates of laminated, varved or
layered clays with particular reference to sand drains. Geotechnique, 14, 321–40.
Rowe, R.K., MacLean, M.D. and Barsvary, A.K. (1984) The observed behaviour of a
geotextile-reinforced embankment constructed on peat. Canadian Geotechnical
Journal, 21, 289–304.
Ruffles, N.J. (1965) Derwent Reservoir. Journal Institution Water Engineers, 19, 361–76.
Saito, A. (1977) Characteristics of penetration resistance of a reclaimed sandy deposit and
their change through vibratory compaction. Soils and Foundations, 17. 32–43.
Saitoh, S., Suzuki, Y. and Shirai, K. (1985) Hardening of soil improved by deep mixing
method. Proc. 11t h International Conf. on Soil Mechanics and Foundation
Samol, H. and Priebel, H. (1985) Soil fracturing—an injection method for ground
movement. Proc. 11th International Conf. on Soil Mechanics and Foundation
Sanger, F.J. (1968) Ground freezing in construction. Proc. American Society Civil
Sanger, F.J. and Sayles, F.H. (1979) Thermal and rheological computations for artificially
frozen ground construction. Proc. 1st Symp. on Ground Freezing, Bochum (ed.
H.L.Jessberger), Elsevier, Amsterdam, pp. 311–37; (also Engineering Geology, 13).
Sayles, F.H. (1989) State-of-the-art: mechanical properties of frozen soil. Proc. 5th
International Symp. on Ground Freezing (eds R.H.Jones and J.T.Holden), Balkema,
Rotterdam, Vol. 1, pp. 143–66.
Sasaki, H. (1985) Effectiveness and applicability of the methods of foundation
improvement for embankments over peat deposits, in Recent Developments in
Ground Improvement Techniques (eds A.S.Balasubramanian, S.Chandra, D.T.
Bergado et al.), Balkema, Rotterdam, pp. 543–62.
Schlosser, F. (1987) Reinforced Earth, in Ground Engineer’s Reference Book (ed. F.
G.Bell), Butterworths, London, pp. 35/1–35/15.
Schlosser, F. (1990) Mechanically stabilized earth retaining structures in Europe, in
Design and Performance of Earth Retaining Structures (eds P.C.Lambe and L.
A.Hansen), American Society Civil Engineers, Geotechnical Special Publication No.
25, pp. 347–78.
Schlosser, F. and Guilloux, A. (1981) Le frottement dans le reinforcement des sols. Revue
Francaise de Geotechnique, 18, 66–71.
Schweitzer, F. (1989) Strength and permeability of single-phase diaphragm walls. Proc.
12th International Conf. on Soil Mechanics and Foundation Engineering, Rio de
Janiero, Vol. 3, pp. 1515–18.
Sherard, J.L., Dunnigan, L.P. and Dekker, R.S. (1976) Identification and nature of
dispersive soils. Proc. American Society of Civil Engineers, Journal Geotechnical
Engineering Division, 102 (GT4), 287–301.
REFERENCES 283
Sherard, J.L., Dunnigan, L.P. and Talbot, J.R. (1984a) Basic properties of sand and gravel
filters. Proc. American Society Civil Engineers, Journal Geotechnical Engineering
Division, 110 (GT6), 684–700.
Sherard, J.L., Dunnigan, L.P. and Talbot, J.R. (1984b) Filters for silts and clays. Proc.
(GT6), 701–18.
Sherwood, F.T. (1957) The stabilization with cement of weathered and sulphate-bearing
clays. Geotechnique, 7, 179–91.
Shibazaki, M. and Ohta, S. (1982) A unique underpinning of soil solidification utilizing
super-high pressure liquid jet. Grouting in Geotechnical Engineering, Speciality
Conf., New Orleans (ed. W.H.Baker), American Society Civil Engineers, New York,
pp. 680–93.
Shuster, J.A. (1972) Controlled freezing for temporary ground support. Proc. 1st North
American Rapid Excavation and Tunneling Conf., Chicago, Vol. 2, pp. 863–94.
Sliwinski, Z. and Fleming, W.G.K. (1975) Practical considerations affecting the
construction of diaphragm walls. Proc. Conf. on Diaphragm Walls and Anchorages,
Institution of Civil Engineers, London, pp. 1–10.
Skempton, A.W. (1964) The long-term stability of clay slopes. Geotechnique, 14,
77–101.
Skempton, A.W., Schuster, R.L. and Petley, D.J. (1969) Joints and fissures in the London
Clay at Wraysbury and Edgeware. Geotechnique, 19, 205–17.
Skipp, B.O. and Renner, L. (1963) The improvement of the mechanical properties of
sands. Proc. Symp. on Grouts and Drilling Muds in Engineering Practice,
Butterworths, London, pp. 29–35.
Somerville, S.H. (1986) Control of Ground water for Temporary Works, Construction
Industry Research and Information Associations, Report 113, London.
Sowers, G.F. (1973) Settlement of waste disposal fills. Proc. 8th International Conf. on
Soil Mechanics and Foundation Engineering, Moscow, Vol. 2, pp. 207–12.
Sridharan, A., Srinivisa Murthy, B.R., Bindumadhava and Revanasiddappa, K. (1991)
Technique for using fine grained soil in reinforced earth. Proc. American Society Civil
Engineers, J. Geotechnical Engineering Division, 117 (8), 1174–90.
Stilley, A.N. (1982) Compaction grouting for foundation stabilization, in Grouting in
Geotechnical Engineering, Speciality Conf., New Orleans (ed. W.H. Baker),
American Society Civil Engineers, New York, pp. 923–37.
Stocker, M.F. and Zwicker, C.A. (1987) Soil improvement by high pressure injection for
the foundation of a highrise building in Kuala Lampur. Proc. 9th South East Asian
Geotechnical Conf., Bangkok, pp. 8/113–8/138.
Talbot, J.R. (1991) Discussion: Improved filter criterion for cohesionless soils. Proc.
American Society Civil Engineers, Journal Geotechnical Engineering Division, 117,
1633–4.
Tan, D.Y. and Clough, G.W. (1980) Ground control for shallow tunnels by soil grouting.
Proc. American Society of Civil Engineers, Journal Geotechnical Engineering
Division, 106 (GT9), 1037–57.
Tang, Y. and Gao, Z. (1989) Experimental study and application of vacuum preloading
for consolidating soft soil foundation. Proc. 12th International Conf. on Soil
Mechanics and Foundation Engineering, Rio de Janiero, Vol. 2, pp. 1423–6.
Tausch, N. (1985) A special grouting method to construct horizontal membranes. Proc.
International Symp. on Recent Developments in Ground Improvement Techniques,
284 REFERENCES
Bangkok (eds A.S.Balasubramanian, S.Chandra, D.T.Bergado, Younger, J.S. and

Prinzl, F.), Balkema, Rotterdam, pp. 351–62.
Temporal, J., Craig, A.H., Harris, D.H. and Brady, K.C. (1989) The use of locally
available fills for reinforced and anchored earth . Proc. 12th International Conf. on
Soil Mechanics and Foundation Engineering, Rio de Janeiro, Vol. 2, pp. 1315–20.
Terzaghi, K. (1925) Erdbaumechanik auf Bodenphysikalischer Grundlage, Deuticke,
Wien.
Terzaghi, K. (1943) Soil Mechanics, Wiley, New York.
Thorburn, S. (1975) Building structures supported by stabilized ground. Geotechnique, 25,
83–94.
Tomlinson, M.J. (1986) Foundation Design and Construction, 5th edn, Longmans,
London.
Toombes, A.F. (1969) The Performance of an Aveling Barford VP 7.7 Mg Self-propelled
Vibrating Roller in the Compaction of Soil, Transport and Road Research Laboratory,
Report LR257, Crowthorne.
Tsytovich, N.A., Abelev, M.Y. and Takhirov, I.G. (1971) Compacting saturated loess
soils by means of lime piles. Proc. 4th European Conf. on Soil Mechanics and
foundation Engineering, Budapest, pp. 837–42.
Van der Merwe, C.J. and Horak, E. (1989) Evaluation of soil/geotextile compatibility.
Proc. 12th International Conf. on Soil Mechanics and Foundation Engineering, Rio
de Janeiro, Vol. 3, pp. 1671–7.
Further reading
General
American Society of Civil Engineers (1987) Soil Improvement: A Ten Year Update,
Geotechnical Engineering Division, Geotechnical Special Publication No. 12, New
York.
Balasubramaniam, A.S., Chandra, S., Bergado, D.T., Younger, J.S. and Prinzl, F. (eds)
(1985) Recent Developments in Ground Improvement Techniques, Balkema,
Rotterdam.
Bell, F.G. (1975) Methods of Treatment of Unstable Ground, Newnes-Butterworths,
London.
Bell, F.G. (ed.) (1987) Ground Engineer’s Reference Book. Part 3, Treatment of the
Ground, Butterworths, London.
British Standards Institution (1981) Code of Practice on Earthworks, BS 6031, London.
David, J.S. (ed.) (1970) Ground Engineering, Institution of Civil Engineers, Thomas
Telford Press, London.
Greenwood, D.A. and Thomson, G.H. (1984) Ground Stabilization: Deep Compaction
and Grouting, Institution of Civil Engineers, Thomas Telford Press, London.
Hausmann, M.R. (1990) Engineering Principles of Ground Modification, McGrawHill,
New York.
Institution of Civil Engineers, (1983) Advances in Piling and Ground Treatment, Thomas
Rathmayer, H.G. and Saari, K.H.O. (eds) (1983) Improvement of ground, Proc. 8th
European Conf. on Soil Mechanics and Foundation Engineering, Helsinki, Balkema,
Rotterdam (3 volumes).
Van Impe, W.F. (1989) Soil Improvement Techniques and Their Evolution, Balkema,
Rotterdam.
Exclusion techniques
Hajnal, I., Martin, J. and Regele, Z. (1984) Construction of Diaphragm Walls, Wiley
Institution of Civil Engineers, (1975) Diaphragm Walls and Anchorages, Thomas Telford
Press, London.
Xanthakos, P.P. (1979) Slurry Walls, McGraw-Hill, New York.
286 FURTHER READING
Ground freezing
Frivik, P.E., Janbu, H., Saetersdal, R. and Finerud, L.I. (eds) (1981) Ground Freezing:
Developments in Geotechnical Engineering, No. 28, Elsevier, Amsterdam (also
Engineering G eology, 18.)
Jessberger, H. (ed.) (1979) Ground Freezing: Developments in Geotechnical
Engineering, No. 26, Elsevier, Amsterdam (also Engineering Geology, 13.)
Chamberlain, E.G. (ed.) (1982) Ground Freezing, Cold Regions, Research and
Engineering Laboratory, Special Report 82–16, Hanover, New Hampshire.
Kinosita, S. and Fukuda, M. (eds) (1985) Ground Freezing, Balkema, Rotterdam.
Jones, R.H. and Holden, T. (eds) (1989) Ground Freezing ’88, Balkema, Rotterdam (2
volumes).
Xiang, Y. and Changsheng, W. (1991) Ground Freezing ’91, Balkema, Rotterdam (2
volumes).
Drainage and dewatering
Cedergren, H. (1986) Seepage, Drainage and Flow Nets, 3rd edn, Wiley, New York.
Holtz, R.D., Jamiolkowski, M., Lancelotta, R. and Pedroni, S. (1992) Performance of
Prefabricated Band Shaped Drains, Construction Industry Research and Information
Association, London.
Institution of Civil Engineers, (1982) Vertical Drains, Thomas Telford Press, London.
Powers, J.P. (1981) Construction Dewatering: A Guide to Theory and Practice, Wiley
Somerville, S.H. (1988) Control of Groundwater for Temporary Works, Report 113,
Construction Industry Research and Information Association, London.
Compaction techniques
Institution of Civil Engineers, (1976) Ground Improvement by Deep Compaction, Thomas

Institution of Civil Engineers, (1978) Clay Fills, Thomas Telford Press, London.
Stamatopoulos, A.C. and Kotzias, P.C. (1985) Soil Improvement by Preloading, Wiley
Soil reinforcement
American Society of Civil Engineers, (1978) Symposium on Earth Reinforcement,

Pittsburgh, New York.
British Standards Institution, (1991) Code of Practice for Strengthened/Reinforced Soils
and Other Fills, BS 8006, London.
Christopher B.R., Gill, S.A., Giroud, J.P., Juran, I., Schlosser, F., Mitchell, J.K. and
Donnicliff, J., (1991) Reinforced Soil Structures. Vol. 1: Design and Construction
Guidelines, Report No. FHWA-RD-89–043, Federal Highways Administration,
Washington, D.C.
FURTHER READING 287
Hanna, T.H. (1982) Foundations in Tension: Ground Anchors, Trans. Tech. Publications,
Clausthal.
Hobst, L. and Zajic, J. (1983) Anchoring in Rock and Soil, 2nd edn, Elsevier, Amsterdam.
Ingold, T.S. (1982) Reinforced Earth, Institution of Civil Engineers, Thomas Telford
Press, London.
Jones, C.J.F.P. (1985) Reinforced Earth, Butterworths, London.
Lambe, P.C. and Hansen, L.A. (eds) (1990) Design and Performance of Earth Retaining
Structures, American Society of Civil Engineers, Geotechnical Special Publication
No. 25, New York.
McGown, A., Yeo, K.C. and Andrawes, K.Z. (eds) (1991) Performance of Reinforced
Soil Structures, Thomas Telford Press, London.
Mitchell, J.K. and Villet, C.B. (eds) (1987) Reinforcement of Earth Slopes and
Embankments, National Cooperative Highway Research Program Report 290,
Transportation Research Board, Washington, D.C.
Shercliff, D.A. (ed.) (1989) Reinforced Embankments: Theory and Practice, Thomas
Technical Memorandum BE3/78, (1987) Reinforced Earth Retaining Walls and Bridge
Abutments for Embankments, Department of Transport, HMSO, London.
Yamanouchi, T., Miara, N. and Ochiai, H. (eds) (1988) Theory and Practice of Earth
Reinforcement, Balkema, Rotterdam.
Geosynthetics
Anon. (1984) A Geotextile Design Guide, Low Bros, Dundee.

Den Hoedt, G. (ed.) (1990) Geotextiles, Geomembranes and Related Products, Balkema,
Rotterdam (3 volumes).
Holtz, R.D. (ed.) (1988) Geosynthetics for Soil Improvement, American Society of Civil
Engineers, Geotechnical Special Publication No. 18. New York.
Ingold, T.S. and Miller, K.S. (1988) Geotextiles Handbook, Thomas Telford Press,
London.
Institution of Civil Engineers, (1984) Proc. of Symposium on Polymer Grid
Reinforcements, Thomas Telford Press, London.
Jarrett, P.M. and McGown, A. (eds) (1987) The Application of Polymeric Reinforcement
in Soil Retaining Structures, NATO Advanced Study Institute Series, Kluwer,
Holland.
John N.W.M. (1987) Geotextiles, Blackie, Glasgow.
Koener, R.M. (1986) Designing with Geotextiles, Prentice-Hall, Englewood Cliffs, New
Jersey.
Koerner, R.M. and Welsh, J.P. (1980) Construction and Geotechnical Engineering using
Synthetic Fabrics, Wiley, New York.
Rankilor, P.R. (1981) Membranes in Ground Engineering, Wiley, Chichester.
Van Zanten, R.V. (1986) Geotextiles and Geomembranes in Civil Engineering, Wiley,
New York.
288 FURTHER READING
Grouting
Baker, W.H. (ed.) (1982) Grouting in Geotechnical Engineering. Speciality Conference,

New Orleans, American Society of Civil Engineers, New York.
Bowen, R. (1981) Grouting in Engineering Practice, 2nd edn, Applied Science, London.
Cambefort, H. (1967) Injection des Sols, Eyrolles, Paris.
Engineering Manual No. 1110–2–3506, (1984) Grouting Technology, US Corps of
Engineers, Washington, D.C.
Houlsby, A.C. (1990) Construction and Design of Cement Grouting, Wiley Inter-science,
New York.
Karol, R.H. (1983) Chemical Grouting, Marcel Dekker, New York.
Nonveiller, E. (1989) Grouting Theory and Practice, Elsevier, Amsterdam.
Verfel, J. (1989) Rock Grouting and Diaphragm Wall Construction, Elsevier,
Amsterdam.
Soil stabilization
Imperial Chemical Industries, (1986) Lime Stabilization Manual, Buxton, Derbyshire.

Ingles, O.G. and Metcalf, J.B. (1977) Soil Stabilization, 2nd edn, Butterworths, Sydney.
National Lime Association, (1982) Lime Stabilization Construction Manual, 7th edn,
Washington, D.C.
Winterhorn, H.F. and Schanid, W.E. (1971) Soil Stabilization Parameters, Technical
Report No. AFWL-TR-70–35, Kirtland Air Force Base, New Mexico.
Site investigation
British Standards Institution, (1981) Code of Practice on Site Investigation, BS 5930,

London.
Clayton, C.R., Simons, N.E. and Matthews, M.C. (1982) Site Investigation, Granada,
London.
Hawkins, A.B. (ed) (1986) Site Investigation Practice: Assessing BS 5930, Engineering
Geology Special Publication No. 2, The Geological Society, London.
Hvorslev, M.J. (1951) Subsurface Exploration and Sampling of Soils for Civil
Engineering Purposes, Waterways Experimental Station, Vicksburg, Mississippi.
Weltman, A.J. and Head, J.M. (1983) Site Investigation Manual, Construction Industry
Research and Information Association, Special Publication 25, London.
Soil types
Bell, F.G. (1992) Engineering Properties of Soils and Rocks, 3rd edn, Butterworth-
Heinemann, Oxford.
Geotechnical Control Office, (1986) Guide to Rock and Soil Descriptions, Geoguide 3,
Government Publication Centre, Hong Kong.
Gillott, J.E. (1987) Clay in Engineering Geology, 2nd edn, Elsevier, Amsterdam.
Mitchell, J.K. (1976) Fundamentals of Soil Behavior, Wiley, New York.
FURTHER READING 289
Young, R.N. and Warkentin, B.P. (1975) Soil Properties and Behaviour, Elsevier,
Amsterdam.
Index
Page numbers in bold type refer to illustrations.

Ablation till 22 in grouts 206–7, 209, 222
Abrasion resistance, geosynthetics 187 slurry 49–52
Accelerators, in grouts 221 sodium 223
Acoustic emission tests 3 wellpoints 96
Acrylamide grouts 216, 228 Bentonite-cement
Activated grouts 219 grouts 220
Active freezing 55 piles 35
Admixtures, for grouts 219–21 Bingham fluids 202
Aggregate filters 68 Biodegradation, of geotextiles 200
Aggregates, cohesionless 12 Bitumen
Alluvial deposits, drainage 84 cutbacks 263–1
Aluminium powder, in grouts 208 emulsions 228–9, 244, 263–4
Aluminous soils 18 Black clays 20
Aminoplast grouts 228 Blasthole spacing 151
Anchored earth 164–6 Bleed rates, cement 208
Anistropy, geosynthetics 184 Bleeding, grouts 206–7, 217, 220–1
Apparent viscosity 204 Blow-outs, deep wells 104
Applied stress 13 Bored wells 102–5
Aquifers, confined 105 Boulder clay 22
Artesian pressure 42, 102 Boulders 5
Augers, deep wells 102 Brines, freezing 55–6
Aureole grouting 239 British Soil Classification 3, 5, 8–11
Autograder 245 Building damage, dynamic compaction 144
Bursting load 185
Backfill, slurry trenches 40
Band drains 84–8, 128 Calcareous clays 21
Barite 43 Calcite, in clays 15
Base exchange 107, 110 Calcium
Basement excavation 106 chloride
Beam method, slurry trenches 41 in electro-osmosis 108
Belling 176 in grouts 221, 224
Bentonite as refrigerant 55, 61
cut-offs 38 in soil stabilization 265
diaphragm walls 43, 45 hydroxide 251 Calcium (cont’d)
290
INDEX 291
silicate, in lime stabilization 252 cement-stabilized soils 244, 247

Calcrete 21 curves 112
Caliche deposits 21 lime addition 254–5
Carbon dioxide, in ground freezing 58 equipment 113–16
Casings grouting 232–3
bored piles 34–5 lime-stabilized soils 257–8
ejector systems 100 mechanical 111–21
micropiles 171–2 methods 117, 120
sand drains 83 piles 128–31
soil anchors 175 soil types 114, 116–21
Cathode well system 107 tests 112–13
Cations, in clay minerals 110 Compressed air cut-offs 52
Cement Compresssibility, of compacted soils 118
content, compressive strength 248 Compression, secondary 124
grouts 208–9, 213–15, 217–21 Compressive strength
hardening rates 210 frozen soils 62–5
in loess 13 sand-silicate mortars 227
microfine 202 Concertina method, face construction 195
requirements, soil types 241 Concrete
stabilization 240–51 columns 141–2
Cement-bentonite cut-off 38 panels, reinforced earth dam 159
Cement-clay grouts 221–3 Cone penetrometer testing 3
Cement-stabilized soils, properties 244–9 Cones of depression
Cement/water ratios, bentonite 37–8 vibroflots 134
Chemical grouts 210–11, 215, 223–8 wellpoints 92
Chlorite 242, 252 wells 105
Chrome-lignin 223, 228, 265 Consistency, grouts 206
Claquage grouting 232 Consolidation
Clay curves 80, 123
balls 242 in electro-osmosis 109
deposits 15–18 primary 13, 16
grouts 209, 222 sand drains 83
soils, expansiveness 16 secondary 77
Clay-chemical grouts 223 testing 3
Cliff protection 201 thaw 67
Coarse-grained soils 12 times, and drain parameters 82
Cobbles 12 Contiguous bored piles 34–5
Coefficient of consolidation 79 Corrosion, geosynthetics 187
Cofferdams 33–4 Counterfort drains 75
Cohesive soils Cratering, explosive compaction 149
compaction 112 Creep
stabilization 241 chemical grouts 211
Colliery discard 27–8 curves, silty clay 66
Colloidal dowelling treatment 167
fine cement 215 in ground freezing 65–6
grouts 219 strength, frozen ground 64
Column jet grouting 235 Creep-relaxation, geosynthetics 186
Compaction Crosshole radar testing 3
292 INDEX
Curing, lime-stabilized soils 255, 259 Dowels 173

Curing time, compressive strength 248 Drain patterns 80
Cushioning, geosynthetics in 189 Drainage
Cut-off techniques 29–30 blankets 73, 77, 84, 199
Cut-offs ditches 74
by grouting 239 fabrics 191
geosynthetics in 199 gallery 76
in slope drainage 75 in reinforced earth 170
Cutback bitumen 263–4 Drilling methods
deep wells 104
Dam cutoffs 2 wellpoints 94
Dams, reinforced earth 159–60 Drop-hammers, in compaction 130
Deformation, cut-off walls 39 Dry density, and compaction 111–13, 119
Densification Durability, of grouts 213
by vibration 120–1, 131, 137 Duricrust 21
compaction piles 129 Dynamic compaction 3, 143–9, 144
particle sizes 134
Density Earthquake
bentonite slurry 52 loadings 120
bulk 13 resistance, reinforced earth 161
dry 13 Earthquakes, soil nailing 170
loess types 14–15 Ejector pumps 97–101
sand 133 Electro-chemical stabilization 108, 110
Depth limits, explosive compaction 151 Electro-osmosis 106–10
Derwent dam 78 Electrode spacing 108
Desiccation cracks 20 Electrolytes, grouts 207
Dewatering 2 Embankments
concentric 97 geogrids in 195–6
systems 77 lime stabilization 259–60
soil types for 93 reinforced soil 197
Diagenetic bonds, clay soils 16 soil-cement in 250
Diaphragm walls 42–9 vibratory rollers 115
Differential settlement Environwall 40
claquage grouting 232 Epoxy grouts 211, 226
jet grouting 235 Erosion control, geosynthetics in 199
peat 25 Eskers 23
sand density 12 ESP, see Exchangeable sodium percentage
sand drains 84 Ethyl acetate, in grouts 225
Dilatancy 13 Evaporites 61
Direct displacement injection 235 Excavation
Discharge capacity tests 86 deep wells 102
Dispersive techniques 39, 47–8
clay 20 Exchangeable sodium percentage 20–1
soils 20–1 Expansive clays, compacted 118
Dissolution, minerals 1 Explosive compaction 149–53
Ditch drainage 74
Double liner systems 200 Fabric-wrapped drains 191
INDEX 293
Face stabilization 169 functions 189–92

Facings 195 properties 181–9
Factors of safety, partial 185 Geotextiles
Fast gel times 229 in filter drains 72–3
Felts 180 properties 179
Ferruginous soils 18, 20 Geowebs 181, 196
Fibre drains 85 Gibbsite 242
Filament reinforcement 157 Goethite 18
Filaments, extruded 180 Grading 5
Fills 25–7, 127 triangle 4
Filter Grain
cake 37–40, 45, 47, 49, 52 packing 13
drains 68, 198 roundness, silts 13
fibres 84 sizes
Filtering, grouts 214 dynamic compaction 147
Filters and grout permeation 203
geosynthetics in 190, 199 Granular fill, reinforcement 157–8, 161
wellpoints 94 Grid reinforcement 156–7
Fin drains 73, 199 Ground
Fines content, penetration resistance 132 anchors 34
Fissured clay 18, 109 freezing 53–67
Flexible drains 76 heating, in electro-osmosis 109
Flocculation 20–1, 222–3, 257 heave 1
Flooding, explosive compaction 152–3 Groundwater
Fluidity, grouts 206 flow 1
Fracture grouting 232 in ground freezing 54, 56, 61
Franki technique 131 heat content 60
Freeze probes 56, 61 pollution 37, 40
Freezing 2 recharge 105
times 59 Grout curtains 239
French drains 199 Grout-hole sizes 218
Friction, reinforced earth 161 Groutability
Frictional resistance, frozen soils 62 limits, soil types 216
Frost ratio 215
action, on compacted soils 119 Grouting 2, 229–30
heave, in ground freezing 66 bored piles 35
sheet piles 33
Gel times, grouts 209, 222, 224–5 soil anchors 174–7
Geocell mattresses 197 soil nailing 170
Geocomposites 181, 194, 199 Grouting pressure 229–30, 233
Geodrain 147 Grouts
Geogrids 154, 181, 189–90, 193, 195–6, chemical 202
198 deterioration 214
Geomats 196, 199 particulate 202, 216
Geomembranes 40, 190, 193, 199 properties 204–14
Geonets 181 solution 202
Geosynthetics suspension 202
Guide walls 46
294 INDEX
Gypsum Kames 23
in arid soils 21 Kaolinite 15, 242, 244, 252, 254, 265
in stabilization 243, 252
Laminated soils, radial flow in 78
Haematite 18 Laterites 18–20, 242
Head, loss of 68, 95 grading curves 19
Heave stabilization 265
clay soils 16 Latosols 20
deep wells 104 Layer clays 15
in grouting 232 Lignin, in soil stabilization 265
Herringbone ditches 74 see also Chrome-lignin
Heterofil bonding 180 Lime
High early strength cement 215 application rates 259
Hole-punching 94–5 columns 88, 260–1, 261
Homofil bonding 180 slurries 257–9
Horizontal slurry injection 260
drains stabilization 251–62
road construction 194 Limestone, nodules 21
in slope drainage 75 Linear drains 73, 199
grouting 236–7 Liquefaction
wellpoints 97–8 dynamic compaction 145
Hurpinoise system 169 granular soils 120
Hydration vibroflotation 131–2, 140
bentonite slurry 50 Liquid limit
minerals 1 black clays 20
Hydraulic loess 14
conductivity, geosynthetics 183 varved clay 24
efficiency, deep wells 102 Load transfer, reinforced earth 161–3
flow, in filter drains 73 Loading tests, and 139
gradients, grouts 211 Lodgement till 22
Hydrofracture 176, 215 Loess
Hydrogels 204 explosive compaction 152
Hydrogeology 60–2 inundation 128
Hydrostatic excess pressure 16 jet grouting 235
particle sizes 14
Illite 15, 21, 242, 244, 252 properties 13–15
Impact energy, and depth 145 soil-cement 250
Instantaneous strength, frozen ground 65 thermal stabilization 267
Interceptor ditches 74, 76 Loma Prieta earthquake 171
Ion transfer 107, 110 London Clay 18, 47
Irrigation canals 251
Made ground 141
Jet grouting 233–7 Marine clays 46
Jet-bolts 170 Mats 181
Jetting 90, 94, 96, 100, 140 biodegradable 200
Joosten process 224 Melt-bonded fibres 180
Membrane-encapsulated soil layer 195
INDEX 295
Menard Packers, in grouting 231

Giga Machine 146 Panel jet grouting 235
pressuremeter 145 Partial factors of safety 185
Metastability, loess 13 Particle sizes
Microbial damage, in classification 3
geosynthetics 187–8 and densification 134
Micropiles filters 70–1
reticulated 171–3 in grouting 214
tunnel protection 172 Particle-retention, geosynthetics in 190
Milchbruck tunnel 66–7 Passive freezing 55
Mine reclamation 198 Peak strength
Mixing clay soils 17–18
cement grouts 219 geosynthetics 184
lime stabilization 253 Peat soils 24–5
Moisture content precompression 125
black clays 20 vibrodisplacement 143
clays 15, 18 Penetration
and compaction 111–13 chemical grouts 216
in ground freezing 59, 61, 64 grouts 214–17
lime-stabilized soils 257 testing 3
loess 14–15 Penetrometer tests 203
peat 24–5 Peptization 207
reinforced earth 163 Permeability 1, 3
Montmorillonite 15, 20, 222, 242, 244, 252, cement-stabilized soils 244, 249
254, 265 of compacted soils 119
Morainic soils, grading 23 electro-osmotic 108
Mullen burst test 185–6 geotextiles 73
lime-stabilized soils 257
Needle-punching 180 loess 15
Newtonian fluids 202, 207, 223, 229 ratio, and sand drains 79
Nitrogen, liquid 58–9, 62 Permeation grouting 230
Non-woven fabrics 180, 191 PFA, see Pulverized fly ash
Nylon 178, 187 pH, bentonite slurry 51
Phenoplast resins 227
Oil tank farms 128 Phosphates
One-shot grouting 225 in bentonite slurry 51
Ordinary Portland cement 210, 215, 217, in grouts 223
240, 243, 263 Phosphoric acid, in soil stabilization 264
Outwash fans 23 Piezometric pressures 29
Overbreak 34 Pipe trenching 98
Overburden pressure, in ground freezing Pipes, in filter drains 73
65, 67 Piping 1, 20, 29, 88, 199
Overcompaction 120 ratio 68–9
Overconsolidation Plastic limit
clay soils 17–18 lime addition 253–4
till 22 varved clay 24
Plastic soil-cement 250–1
Plasticity, clays 15, 18
296 INDEX
Plasticity index Pulverization

black clays 20 cement-stabilized soils 241, 243, 245
lime-stabilized soils 253–4 lime-stabilized soils 257
loess 14 Pulverized fly ash 47, 220, 252, 262–3
till 22 Pumps
Polyamides, see Nylon submersible 104
Polyester grouts 211, 226–7 sump 89
Polyesters, geosynthetics 178, 186 wellpoints 92
Polyethylene 178, 181 PVC 178
Polymer grid materials 156–7 Pyrite, in colliery discard 28
Polypropylene 178, 181, 186
Polyvinylidene chloride, see PVC Quarry tailings ponds 198
Pore pressure Quick clays 24
colliery discard 28 Quick conditions 1, 29
in ground freezing 65 Quicklime 251, 257, 262
Pore size, geosynthetics 182
Pore-water Radial drainage 77–9
in clay soils 16 Rammers 115
pressure Rapid-hardening cement 240
clay soils 17–18 Red earths 20
in compaction 111, 118, 120 Refrigerants
and consolidation 126 expendable 57–8
in electro-osmosis 109 secondary 61–2
in grouting 233 Refrigeration
precompression 125–6 capacity 59
sand drains 80 techniques 55
vibrocompaction 138 Reinforced earth 154–66
Porosity, silts 13 dams 160
Post-construction movement, cohesive uses of 161
soils 164 Reinforcement
Pot life, grouts 224, 227 geosynthetics in 190
Potential gradients 108 three-dimensional 157
Pounders 143–4 Reinforcing materials 156
Pozzolans 220, 252, 263 Relative compaction 112
Precompression 121–8 Relief wells 105
surcharge loading 122 Remoulding, clay soils 17
Preloading 121 Replacement
tests 3 fills 250
Preponding 128 jet grouting 233–5
Pressuremeter tests 203 Reservoirs, soil-cement in 250
dynamic compaction 148 Residual strength, clay soils 17–18
Primary settlement, explosive compaction Resin grouts 226–7
152 Resin-bonded fabrics 181, 191
Proctor test 112 Resorcinol resin 227–8
Pull-out Retaining walls, reinforced earth 161
resistance 164 Retarded cement 240
soil anchors 174 Reverse grouting 236–7
tests, reinforced earth 156
INDEX 297
Rheopexy 209 till 23

Rigidity times, fills 26
geosynthetics 182 Shaft sinking 53
grouts 204, 217 ground freezing 57
River defence works 200 Shale, expanded 252
Road construction, geosynthetics in 192–4 Shallow piles, plastic soil-cement in 251
Rollers Shear
rubber-tyred 113, 243 failure, sand drains 83
sheepsfoot 113 resistance, clay soils 17
smooth-wheeled 115 strength
vibratory 115, 119–20, 243 cement grouts 208
Root piles 171 cement pastes 205
Root-holes, in loess 13, 15 clay grouts 209
Rotary tillers 243 clays 15
Rupture strength, geosynthetics 184 colliery discard 28
grouts 210
Sand peat 25
columns, wellpoints 96 reinforced earth 162, 164
drains 77–80, 83, 105, 127, 262 Shearing
piles 129–31 displacement, vibration compaction
Sandbags, earth anchors 166 142
Sandwicks 84, 127 mechanisms, earth anchors 165
Sanitary landfills 27, 250 resistance, loess 14–15
Scaling, ejector pumps 100 Sheet piles 31–3
Scarification 257 recharge wells 106
Sea defences 200 Shot nails 167
Secant piles 35–7, 36 Shotcrete 167, 169–70
Secondary Shrinkage
compression, surchage 124 cement-stabilized soils 245
consolidation, fills 27 grouts 211
Seismic testing 3 Shrinkage limit
Self-healing layers, in filters 71 black clays 20
Sensitivity, clays 18 lime-stabilized soils 254
Separation, geosynthetics in 189 Silicate grouts 202, 210
Serre Poncon dam 238 Silt, consolidation 13
Setting Site
agents, grouts 222 history 2
times, resin grouts 227 investigation 2–3
Settlement Sizewell power station 47
after ground freezing 67 Slag, in lime stabilization 252
in compacted soils 118 Sleeve-grout 231
construction 13 Sliding method, face construction 195–6
differential 12, 84, 232 Slope
fill 26 Settlement (cont’d) drainage 74–6
preloading 121 failure, artesian pressures 29
reinforced earth 195 reinstatement 198
through groundwater abstraction 105 stability, wellpoint 93
stabilization 260
298 INDEX
micropiles 173 Stone columns, vibrocompaction 138–41

reinforced earth 160 Strength, lime-stabilized soils 254–5
thermal method 267 Strength development, water/cement ratios
Slurry trench cutter 45 212
Slurry trenches 37–40 Stress decay 187
Smear effects 79, 85, 88 Stress-strain
Smectite 21, 110 characteristics, geosynthetics 183, 191–
Sodium 2
aluminate 223 curves, filaments 184
chloride, in soil stabilization 264 relationships, frozen soils 63
hydroxide, in soil stabilization 265 Strutted piling 31
phosphate, in grouts 222 Subgrade stabilization 257
silicate Suction
in electro-osmosis 108 lift, wellpoints 96
in grouts 221–5 wells 94
Soil Sulphates 27–8
anchors 174–7 Sumps 89
classification 3–12 Surcharge 121, 124, 127–8, 137
tables 6–12 Surface runoff 74
deformation 1 Surfactants, in grouts 220
disturbance, sand drains 79 Swell potential 16
dowels 173 Swell-shrink response, clays 15
nailing 167–71 Swelling
reinforcement 2 of compacted soils 118–19
strength 1 capacities, clay minerals 15
types 12–28 period, bentonite slurry 50
and dewatering systems 93 Syneresis 213, 226
reinforced earth 163 Synthetic filters 72
Soil-bentonite cut-offs 38, 40
Soil-cement Tamping 143
cushions 250 Tear propagation 185–6
properties 246 Telescopic method, face construction 195
sub-base 244 Temperatures
uses 250 freezing brines 56–7
Soil-lime, properties 246 for thermal stabilization 265–7
Soldier beams 33 Tensile resistance, geosynthetics 184
Soletanche Hydrofraise rig 48 Tension piles 176
Spacing, wellpoints 95 Terra-probes 137–8
Spoil heaps 28 Terraces, kame 23
Spontaneous combustion, colliery discard Terzaghi criterion 69–70, 72, 122
28 Thermal
Squeezing ground 34 conductivity 54
Stability stabilization 265–7
cement grouts 218 Thixotropic suspensions 206–7
grouts 207 Thixotropy
Stabilizers, additives 240 bentonite grouts 222–3
Stage grouting 230 cement grouts 220
Steel sheet piles 31, 32
INDEX 299
Till 22 grouts 217

Tolerances, diaphragm walls 45 pastes 205
Total cation concentration 21 grouts 204, 235
Tremie process 35, 43, 45, 47, 51–2, 141, Void ratios 12–13
251 clay soils 16
Trench collapse 46 peat 24
Triaxial tests reinforced earth 163
grouted sands 213
reinforced sand 155 Waling 33
Tropical soils 18–22 Wall thickness, ground freezing 59
Tube-à-manchette 176, 230–1, 237, 239 Walls, earth-reinforced 195
Tunnel Waste disposal
driving 53 fills 27
excavation, grouting in 239 sites 199, 250
Tunnels, soil nailing 168 Water
Two-shot grouting 224 in frozen soil 55
content, grout properties 218
Ultra fine cement 215 jetting, band drains 86
Ultraviolet light, geosynthetics 188 movement, under frozen ground 61
Umbrella grouting 239 retention, grouts 206–7
Underground construction, micropiles 172 Water/cement ratios, in grouts 217
Underpinning, by grouting 232–3 Weathering
Underreaming 176 geosynthetics 188
Unified Soil Classification 3, 6–7 latosols 20
Uplift pressures, road construction 194 Welded-wire grids 157
Urea formaldehyde, in soil stabilization Well development 104
265 Wellpoints 2, 41, 43, 90, 90–8
multi-stage 93, 96–7
Vacuum preloading 127–8 spacing 96
Varved clay 23 Wells, bored 102–5
Venturi effects, ejector pumps 100 Wet strength, grouts 210
Vermiculite 242 Wing jet grouting 235
Vertical drains 77, 79, 83 Woven fabrics 178, 180, 182, 184 191
dynamic compaction 147
layout 81 Yield point, grouts 206
in precompression 121 Young’s modulus
preloading 127 cement-stabilized soils 247
Vibration, compaction piles 128–9 reinforced earth 161
Vibro-composer methods 130
Vibrocompaction 131–43
Vibrodisplacement 142–3
Vibroflot 135
Vibrofloatation 3, 131–6
Vibroflots 131, 133–6, 142
Vibroreplacement 138–42
Viscosity
cement

Engineering Treatment of Soils

Uploaded by

Copyright:

Available Formats

Engineering Treatment of Soils

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Engineering Treatment of Soils

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What are some of the topics covered in the book?

What are some of the topics covered in the book?

Engineering Treatment of Soils

OTHER TITLES FROM E & FN SPON

ISBN 0-203-47418-X Master e-book ISBN

ISBN 0-203-78242-9 (Adobe eReader Format)

3.3 Hydrogeology and the nature of frozen soils 59

7 Soil reinforcement and soil anchors 151

9.4.5 Chemical grouts 218

Appendix: Summary of soil-treatment methods 263

The engineering treatment of soils involves improving their geotechnical

of the Geotechnical Engineering Division; and the Proceedings of the

Table 1.1 Symbols used in the Casagrande soil classification

Table 1.5 Description of material/soil mixtures

Boulders (or Cobbles) with a little Up to 5% finer material

Gravels Sands Silts

1. Peak strength of intact clay: c’=31 kPa, '=20;

The greatest variation in the engineering properties of clays can be attributed to

Figure 1.5 Grading curves of laterites (after Madu, 1977)

Table 1.7 Some common properties of laterites

material usually contributes up to about 80% of basal or lodgement tills.

Material Characteristics as fill

Groundwater frequently provides one of the most difficult problems during

the existence of an impermeable stratum beneath the excavation to form an

Contiguous piles can be designed to cantilever vertically, to be strutted or to

Where water cannot be removed from a borehole either by baling or pumping,

high-density polyethylene sheeting can be as low as 10−14 m/s. When a slurry

inadequate concreting could cause disastrous flooding in deep excavations. This

parks. Furthermore, the technique is virtually vibrationless and quiet, and so

Figure 2.6 Construction of a reinforced diaphragm wall. General procedure for

If excessive loss of bentonite occurs in pervious strata, especially if the

Figure 2.8 Diagrammatic outline of the Soletanche Hydrofraise rig

1. support the excavation by exerting hydrostatic pressure on its walls;

The pH value is particularly significant where the chemistry of the soil

the effects become noticeable at 0.3% cement contamination, which is the

Figure 3.1 Thermal conductivity of soil (from Mitchell, 1981)

Figure 3.3 Ground freezing with supercooled brine

Figure 3.4 Ground freezing layout for an excavation

Figure 3.5 Ground freezing with liquid nitrogen

temperatures are required, then alternative secondary refrigerants or circulation of

Figure 3.10 Compressive strength of frozen sand and clay

ground by overburden pressures or any applied loads. Excess pore water

was conservative. It can be considered to have a safety factor of about 2. Thus

as the principal filter acceptance criterion. This conclusion applies generally to

Figure 4.2 Linear drains enclosed with geotextile

distribution, shape and packing) in which it is incorporated, and the hydraulic

where Tr is the dimensionless time factor for radial consolidation, D is the

Table 4.2 Main characteristics of some common types of band drains

Data Units Geodrain Geodrain KB Desol

Alidrain Mebradrain MD 7007 Colbond CX 1000 PVC

Alidrain Mebradrain MD 7007 Colbond CX 1000 PVC

Figure 4.16 Placement of band drains

Dewatering is one type of temporary measure taken to lower the groundwater

Figure 5.1 Wellpoint system dewatering an excavation

Figure 5.4 Multi-stage wellpoint system

Figure 5.5 Soil types and dewatering systems

Figure 5.6 Hole-puncher and casing

wellpoint installed before the hole-puncher is extracted. When equipped with a