Underwater embankments

on soft soil: a case history

BALKEMA Proceedings and Monographs
in Engineering, Water and Earth Sciences
Underwater embankments
on soft soil: a case history
W.F. Van Impe
R.D. Verastegui Flores
University of Ghent, Laboratory of Geotechnics,
Zwijnaarde, Belgium
LONDON / LEIDEN / NEW YORK / PHILADELPHIA / SINGAPORE
Taylor & Francis is an imprint of the Taylor & Francis Group, an informa business
c 2007 Taylor & Francis Group, London, UK
All rights reserved. No part of this publication or the information contained herein
may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, by photocopying, recording or otherwise, without
written prior permission from the publishers.
Although all care is taken to ensure integrity and the quality of this publication and
the information herein, no responsibility is assumed by the publishers nor the author
for any damage to the property or persons as a result of operation or use of this
publication and/or the information contained herein.
Published by: Taylor & Francis/Balkema
P.O. Box 447, 2300 AK Leiden, The Netherlands
e-mail: [Link]@[Link]
[Link], [Link], [Link]
Library of Congress Cataloging-in-Publication Data
Van Impe, W.F.
Underwater embankments on soft soil : a case history / W.F. Van Impe,
R.D. Verastegui Flores.
p. cm.
Includes bibliographical references.
ISBN 978-0-415-42603-9 (hardcover : alk. paper)
1. Embankments. 2. Soil stabilization. 3. Underwater construction. I. Verastegui
Flores, R. D. II. Title.
TA760.V3576 2007
627

.58dc22 2007001786
ISBN: 978-0-415-42603-9(Hbk)
ISBN: 978-0-203-94629-9(eBook)
This edition published in the Taylor & Francis e-Library, 2007.
To purchase your own copy of this or any of Taylor & Francis or Routledges
collection of thousands of eBooks please go to [Link].
ISBN 0-203-94629-4 Master e-book ISBN
Contents
About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI
List of symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Construction on soft soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 Clay foundation behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Staged construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.1 Methodology for analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.2 Consolidation analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3 Discussion on undrained shear strength of soft soils . . . . . . . . 11
3.1 Preamble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Shear behavior of normally consolidated clay . . . . . . . . . . . . . . . . 11
3.3 Normalized behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 Shear behavior of overconsolidated clay . . . . . . . . . . . . . . . . . . . . 14
3.5 Laboratory testing techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.5.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.5.2 SHANSEP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.6 Undrained strength anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.7 Final recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4 Discussion on slope stability evaluation . . . . . . . . . . . . . . . . . . . . 25
4.1 Preamble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.2 Causes of slope instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3 Stability conditions for analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4 Stability analysis procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
VI Contents
4.4.1 Limit equilibrium methods . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.4.2 Strength reduction methods . . . . . . . . . . . . . . . . . . . . . . . . 31
4.4.3 Limit equilibrium versus strength reduction methods . . . 33
4.5 Failure mechanisms for highly sensitive clays . . . . . . . . . . . . . . . . 35
4.5.1 Flake type sliding of quick clay . . . . . . . . . . . . . . . . . . . . . . 37
4.5.2 Analysis taking pore water pressures into account . . . . . 40
4.5.3 Mechanism of sliding in quick clay masses . . . . . . . . . . . . 44
4.5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.6 Risk of liquefaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.7 Slope stability analysis of the Doeldok embankment . . . . . . . . . . 51
5 Evaluation of consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 One-dimensional consolidation theory . . . . . . . . . . . . . . . . . . . . . . 57
5.2 Innitesimal strain theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.2.1 Numerical solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.2.2 Applications of SSCON-FD . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.3 Finite strain theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.4 Innitesimal strain versus nite strain theory . . . . . . . . . . . . . . . 66
5.5 Consolidation at the Doeldok site . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6 Geotechnical characterization of the site . . . . . . . . . . . . . . . . . . . 71
6.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.2 Soil prole and characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.3 Selection of parameters for design . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.3.1 Dredged material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.3.2 Boom clay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.3.3 Sand. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.3.4 Summary of soil properties . . . . . . . . . . . . . . . . . . . . . . . . . 80
7 Design of underwater embankment of soft soil . . . . . . . . . . . . . 81
7.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.2 Geometry of the embankment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.3 Stability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.3.1 Undrained analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.3.2 Drained stability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.4 Settlements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.4.1 From Constitutive relationships . . . . . . . . . . . . . . . . . . . . . 87
7.4.2 From nite element program . . . . . . . . . . . . . . . . . . . . . . . . 88
7.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8 Ground improvement by deep mixing . . . . . . . . . . . . . . . . . . . . . . 91
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
8.2 Deep mixing applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8.3 Mechanism of stabilization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Contents VII
8.3.1 Stabilization with lime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8.3.2 Stabilization with cement-like binders . . . . . . . . . . . . . . . . 94
8.4 Methods of installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
8.5 Belgian experience on on-land deep mixing . . . . . . . . . . . . . . . . . 97
8.5.1 Properties of untreated soils on land . . . . . . . . . . . . . . . . . 98
8.5.2 Binders employed on land . . . . . . . . . . . . . . . . . . . . . . . . . . 100
8.5.3 Lime-cement stabilization in the laboratory . . . . . . . . . . . 100
8.5.4 Lime-cement stabilization in-situ . . . . . . . . . . . . . . . . . . . . 102
8.5.5 Remarks on the experience of dry deep mixing on land . 107
8.6 Deep mixing assessment on the underwater site . . . . . . . . . . . . . 109
8.6.1 Properties of the articially cemented soil in the
laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8.6.2 Properties of the cemented soil in the eld . . . . . . . . . . . . 116
8.6.3 Laboratory versus in-situ behavior . . . . . . . . . . . . . . . . . . . 119
9 Construction and monitoring of embankment . . . . . . . . . . . . . . 123
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
9.2 Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
9.3 Quality control of the embankment sand . . . . . . . . . . . . . . . . . . . 124
9.4 Instrumentation and monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
9.4.1 Excess pore water pressure . . . . . . . . . . . . . . . . . . . . . . . . . 129
9.4.2 Settlements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
9.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
About the Authors
William Frans Van Impe
Professor Van Impe was appointed Professor in the
Faculty of Applied Sciences of Ghent University in
1982 and has established a Laboratory of Soil Me-
chanics which enjoys both a national and interna-
tional reputation. Since 1988, he has been a visiting
professor at the Katholieke Universiteit Leuven and
is member of the Board of Directors of the Royal
Academy of Overseas Sciences and board member
of the European DFI group. From 1994-1998 he was
vice-president of the ISSMGE, Europe and in 2001
became President of the ISSMGE for a four-year term. His research has been
rewarded with several international prizes and awards. He and his team have
authored three books and over 200 papers in journals and proceedings. Pro-
fessor Van Impe has acted as a geotechnical consultant to numerous overseas
projects and was the Geotechnics Committee member of the project evaluation
of the Messina Strait Crossing - the largest suspension bridge in the world.
Ramiro Daniel Verastegui Flores
R. D. Ver astegui Flores was born in Bolivia in 1976.
He graduated as a Civil Engineer in 1999 at San Simon
University (UMSS) in Bolivia where he was awarded a
prize in recognition of his research on soil dynamics
by the Faculty of Technology and Science at UMSS. In
2002 he completed a Post-graduate Degree on Geotech-
nical Engineering at Ghent University in Belgium un-
der the guidance of Prof. W. F. Van Impe. Since then,
he has been involved in research projects dealing with
soft soils engineering in near shore and oshore con-
ditions. Currently he is an Academic Assistant and PhD student at Ghent
University. He is author and co-author of a number of papers in international
conference proceedings and journals. His interests focus on stress-strain be-
haviour of cemented soils, ground improvement techniques, large strain con-
solidation of soft soils and deep foundations.
Preface
Ground improvement is probably the oldest but, from a technical point of
view, still the most intriguing technique of all common execution methods
in foundation engineering. Indeed, ground improvement was already in use
more than 3000 years ago. In recent decades, the modern methods of ground
improvement are making use of explosives, impact energy, thermal treatment
of the soil, vacuum consolidation, vibratory compaction technologies, stabi-
lization and solidication of soft soils, as well as combined systems of ingenious
grouting systems, deep mixing techniques, etc.
Throughout the world, deep mixing techniques today are of utmost impor-
tance in dealing with more and more demanding foundation problems. This
tendency has been noticed in Belgium already at a very early stage; with
inventive new developments of soft soil deep mixing technologies and various
advanced high pressure mixing methods. Some initial experiences onshore and
oshore have proved already some years ago that successful solutions can be
attained.
The present work illustrates a challenging example of design and con-
struction of a quite important large underwater embankment on very soft
soil. Throughout the design staged construction and ground improvement by
deep mixing combined with geotextile reinforcement were proposed to assure
the safety as well as allowable deformations of the construction.
The outcome of monitoring excess pore water pressures and displacements
during the construction shows that when taking account of key aspects of
advanced soil stress-strain behavior, it is possible to appropriately model such
complex problem and even to make rather simple attempts to reach some of
the type A foundation behavior predictions.
We may expect this work to be recognized as a valuable reference case
history for the geotechnical engineer, both from the academic as well as from
the practitioners point of view, in order to contribute to the art of building
on soft soils.
Our acknowledgments go to the contractors DEME and J. DE NUL, to
the teams of geotechnical experts of Dredging International NV, HydroSoil
XII Preface
Services, to the Geotechnical Division and Maritime Access Division of the
Flemish Ministry, all actively contributing to the satisfactory result of this un-
common foundation engineering problem. The expert group following closely
the design and construction of the embankment is listed here:
J. Van Mieghem Flemish Ministry, Department of Maritime Access
H. De Preter Flemish Ministry, Department of Maritime Access
J. Van den Broecke Flemish Ministry, Department of Maritime Access
C. Boone Technum, Belgium
H. Cecat Technum, Belgium
M. Van den Broeck Dredging International
R. Aelvoet Combinatie Kallo
P. Menge Combinatie Kallo
S. Vandycke Combinatie Kallo
F. Verhees Combinatie Kallo
R. Lheureux Combinatie Kallo
P. De Schrijver Flemish Ministry, Department of Geotechnics
A. Baertsoen Flemish Ministry, Department of Geotechnics
R. Simons SECO, Belgium
R. Dedeyne SECO, Belgium
In each of the various chapters of the book, the discussion contributing
expert group members have been mentioned.
Ghent W.F. Van Impe
February 2007 R.D. Ver astegui F.
List of symbols

] shear angle

1
,
3
[kPa] principal stresses

p
[kPa] preconsolidation pressure

v
[kPa] vertical eective stress

v0
[kPa] initial vertical eective stress
CRR cyclic resistance ratio
CSR cyclic stress ratio
c

[kPa] eective cohesion

c
u
[kPa] undrained shear strength
c
v
[m
2
/s] coecient of consolidation
e void ratio
E

[MPa] drained Youngs modulus

E
u
[MPa] undrained Youngs modulus
f
s
[MPa] cone sleeve friction
k [m/s] hydraulic conductivity
K
0
horizontal stress coecient at rest
m
v
[1/kPa] coecient of volume compressibility
p

eective mean stress

q

deviatoric stress
q
c
[MPa] cone penetration pressure
q
u
[kPa, MPa] unconned compressive strength
R shear strength reduction factor
r
k
anisotropic hydraulic conductivity ratio
S undrained strength ratio
u [kPa] pore water pressure
1
Introduction
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
J. Van Mieghem
Maritime Access Division, Ministry of Flanders, Belgium
As in many harbor area all around the world, there is an increasing need of
reclaimed land for storing excavated soil in the harbor of Antwerp in Belgium,
mainly because of the construction of new docks. This fact has encouraged
the partial lling up of a dock with a partially submerged embankment as
a retaining structure. Figure 1.1 illustrates the location of the embankment
in the dock (Doel dock) at the harbor of Antwerp. The design and ongoing
construction of a partially submerged 27-m high sand embankment to be
founded in this case on about 8m of very soft soil (not removable because of
geoenvironmental considerations) created an uncommon challenge.
The soil prole at the site consists of an upper soft layer overlying a thin
tertiary sand and a very thick tertiary overconsolidated clay layer. Chapter
6 shows in more detail the geotechnical characterization of the site. The soft
material is the result of years of sedimentation and self weight consolidation of
dredged material from the harbor waterways. Out of preliminary eld and lab
Figure 1.1. Location on the embankment in the Doel dock
2 Introduction
testing it was found that this dredged material still remained in a NC state
with an initial shear strength of about c
u
=3 kPa. The consolidation behavior
of the soft material was studied with more precision later on through CRS,
oedometer and hydraulic conductivity tests.
Throughout the design phase (see Chapter 7) numerous preliminary design
options were worked out aiming at optimizing the dam design. Given the
soft consistency of the dredged material, it became clear that some kind of
foundation layer reinforcement was needed. Therefore ground improvement
by a novel deep mixing technology (SSI) was proposed, mainly applied at the
toes of the embankment. These improved zones are meant to provide extra
strength and to serve as a means of connement of the soft soil in the middle
part of the embankment foundation layer.
The analysis showed, as expected, that the construction phase was actually
the most critical stage for the embankment stability. Unavoidably, a staged
construction was implemented. Staged construction is a technique that uses
controlled rates of loading to enable soil strength to develop via consolida-
tion in order to increase the embankment stability. Consequently, an accurate
evaluation of the consolidation progress (taking account of large deformations)
had to be the key issue for this foundation problem on very soft soil.
The soft material was improved by deep mixing columns introducing the
SSI technique. This deep mixing technique (SSI) makes use of a combination
of highly pressurized cement slurry jetted through a set of nozzles along the
front side of a rotating arm on the one hand and a series of low pressure
cement slurry nozzles on the back side of the rotating arm on the other hand.
The rotation of the arm is combined with a continuous uplift movement; the
cement mixing in this respect is uniformly distributed along the diameter of
the column.
An extensive laboratory research program was set up to study the behavior
of the improved material (see Chapter 8). A number of cement type were tried
out and strength measurements were performed up to long periods of curing.
Control of actual column strength by means of unconned compression tests
on core specimens showed that the strength reached in the eld was higher
than the strength expected from laboratory prepared specimens. This fact
was also thoroughly explained by looking at the microstructure of specimens
from the laboratory and the eld under the microscope (the scanning electron
microscope).
The behavior of the embankment, still in construction today, has been
assessed by means of elaborated instrumentation to measure excess pore water
pressure (PWP) and settlements of the foundation layer (see Chapter 9). As
expected, the measurements do show a slow dissipation of PWP; the excess
PWP measured within the SSI-column improved zone is by far smaller than
PWP within the untreated zone because of the stiness of the SSI treated soil
columns. As for the vertical displacements and PWP observed so far, they are
in good agreement with the expected values from the design.
2
Construction on soft soil
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
2.1 Clay foundation behavior
Construction on soft soils remains a troublesome foundation engineering chal-
lenge as very often a signicant load is imposed by the new structure and the
available shear strength of the soft clay is low.
Traditionally, the analysis of such problem is made in two steps that
consist of:
Slope stability analysis during construction under fully undrained condi-
tions.
Slope stability analysis long after the end of construction under fully
drained conditions.
The validity of this approach was shown not always satisfactory as reported
already by Leroueil et al. (1990) after they compared actual measurements
and predictions assuming fully undrained conditions. Such observations, for
dierent soil conditions, can be explained looking at the stress path of an
element under the centerline of e.g. an embankment load.
Figure 2.1 illustrates the case of an embankment on slightly overconsoli-
dated (OC) clay. The most likely stress path evaluated from many observa-
tions is depicted by the path ABC. The stress state of the soil starts at A
and reaches the yield surface in a zone (B) where the vertical eective stress

v
is approximately equal to the preconsolidation of the clay

p
. Throughout
path AB, the coecient of consolidation is usually high; then, a relatively fast
dissipation of excess pore water pressure can be expected. However, once the
soil reaches the yield surface it becomes normally consolidated (NC) and a
rather signicant drop of the coecient of consolidation is usually observed.
Therefore, only the BC path may be considered essentially as an undrained
(slow dissipation of PWP).
Similarly, the behavior of a highly OC clay foundation is illustrated in
gure 2.2. The initial state of stress is closer to an isotropic state (see point A).
4 Construction on soft soil
Figure 2.1. Behavior of a slightly OC clay foundation
Figure 2.2. Behavior of a highly OC clay foundation
Observations show that during undrained loading, overconsolidated clays
generate lower pore pressures than soft clays. Their high stiness implies a
high coecient of consolidation. Unlike slightly OC clays, typical stress paths
of highly OC clays reach the yielding surface (B) at a value of vertical eective
stress lower than the preconsolidation pressure. Any further loading follows a
path along the yield surface (BC) where slow dissipation of excess pore water
pressure can actually occur.
As a result, the assumption of fully undrained conditions becomes valid
only for normally consolidated clays (see Fig. 2.3) since their initial stress state
falls already in the yielding surface. In fact, many experiences show trends of
excess pore water pressure measurements such as those illustrated in gure
2.4. In NC clays, the excess pore water pressure increases at about the same
rate as the total applied stress (undrained conditions), however, in OC clays
the pore pressures generated at the start of the construction are low and
then rise roughly at the same rate as the total applied vertical stress once the
embankment has reached a critical loading (related to the yielding of the soil).
In the specic case of the design of the underwater embankment, it was
observed that the foundation soil is almost fully normally consolidated, then
assuming fully undrained conditions for stability analysis is not mistaken.
Staged construction 5
Figure 2.3. Behavior of a NC clay foundation
Figure 2.4. Excess pore water pressure mobilization
Sometimes, as in the present case, it is necessary to go beyond the capacity
of the clay. Here, the total load imposed by the embankment is greater than
the bearing capacity of the foundation soil. To tackle this problem the engineer
is left with two options, either to choose for some kind of ground improvement
technique to increase the soft soil strength or to work with the soil allowing
it to drain and increase its own strength by running the construction at a
controlled rate. This second option is called staged construction and it is
discussed in detail in the next section.
2.2 Staged construction
Staged construction is a technique employed in soft soil construction, where
the imposed loading is suciently large to stress the cohesive foundation soils
beyond their preconsolidation pressure and close to failure. Examples include
6 Construction on soft soil
p'
q'
CSL
Ko-line
e
p'
NCL
CSL
v
1
v
2
v
3
cu
3
cu
2
cu
1
e
1
e
2
e
3
1
2
3
P

C
E

D
F
G
A
C
E
B
D
F
G
1st Loading: AB
Consolidation: BC
2nd Loading: CD
Consolidation: DE
3rd Loading: EF
Consolidation: FG
Figure 2.5. Principle of staged construction
embankments for transportation facilities, ood-control levees, tailing dams,
landlls and oshore structures.
Since these projects generate positive excess pore water pressure within
the foundation soil, the most critical condition occurs during construction;
therefore, undrained stability analysis are commonly implemented. As pre-
viously discussed, the assumption of fully undrained conditions ( = u)
for stability analysis is appropriate only for saturated normally consolidated
clays.
In order to have a better understanding of the behavior of the foundation
soil subjected to a load, gure 2.5 outlines the eective stress path of a nor-
mally consolidated element located under the central line of an embankment
constructed in stages.
Staged construction 7
We can see that the total equivalent load applied (represented by point G)
is greater than the initial bearing capacity of the soft soil foundation; if the
construction would be carried out in a single step, failure would indeed occur.
Then, a staged construction has to be implemented to ensure safety during
construction.
Three stages have been outlined, the eective stress path for the rst
loading stage is AB. Loading is halted before failure occurs. The initially
available undrained strength is c
u1
. If full consolidation is allowed, the
excess pore water pressure dissipates and the eective stress path goes from
B to C. Consolidation leads to a reduction of the void ratio as density
increases.
Note that for the second loading stage (path CD) an increased undrained
strength c
u2
becomes available now. Again, at the end of the second stage
(point D) consolidation is allowed (path DE) and nally, the third stage EF
can be safely constructed. However, the drawback of the staged construction
method relates to the more extended period for construction required. When
construction time is not an issue, the implementation of the staged construc-
tion principle in a project is a cost eective solution.
2.2.1 Methodology for analysis
The design of staged construction projects usually entails the steps sum-
marized in gure 2.6. First it is necessary to evaluate stability for the rst
stage loading assuming no drainage during construction, based on the initial
mechanical parameters of the soil. Then, stability calculations during subse-
quent construction stages can be made taking into account that the combi-
nation of the previously applied loads and either partial or full consolidation
will change the initial stress history and will increase the available strength
of the foundation soil.
In gure 2.6, the initial state variables refer to the soil prole and pre-
consolidation pressure. These, together with laboratory testing results (to
measure undrained strength) supply the basic information.
The rst stability analysis makes use of initial soil data to compute the
factor of safety of stage 1. Subsequent stability evaluations require knowledge
of consolidation degree in order to predict a new undrained strength.
2.2.2 Consolidation analysis
Clearly, staged construction design requires an accurate consolidation analysis
to predict rates of pore water dissipation during construction. These predic-
tions often have a strong impact on project feasibility, schedule and costs
during design.
In the following paragraphs some problems regarding consolidation evalu-
ation are identied.
8 Construction on soft soil
Figure 2.6. Methodology for staged construction analysis
[Link] One-dimensional consolidation
Although practice often relies on the conventional 1D small strain consol-
idation theory after Terzaghi (1925), the assumption of constant values of
coecient of consolidation (c
v
) and compressibility (m
v
) may give poor esti-
mates (non conservative) of pore water pressure dissipation during construc-
tion, especially when dealing with very soft soils in which parameters may
signicantly change with strain. Problems may also arise when dealing with
moderately overconsolidated deposits (which suer large changes in both c
v
and m
v
near the overconsolidation pressure p

) and layered clay deposits hav-

ing dierent c
v
and m
v
values.
Such situations do require a modeling technique that allows the use of non
linear functions to describe the consolidation parameters, changes of load with
time and less restrictive boundary conditions. Chapter 5 examines this topic
in more detail.
Staged construction 9
[Link] Accelerating consolidation with drains
The primary purpose of vertical drains, often implemented together with
staged construction, is to speed up the consolidation process by allowing hor-
izontal drainage and shortening the drainage path as well. In addition, most
natural deposits are anisotropic with respect to ow properties, the horizontal
hydraulic conductivity being typically higher than the vertical one.
Installation of vertical drains to accelerate the rate of consolidation intro-
duces two additional problems regarding reliable predictions of pore pressure
during staged construction: evaluation of the anisotropic permeability ratio
r
k
= k
h
/k
v
, required to estimate the coecient of consolidation for horizontal
ow; and assessment of the likely eects of soil disturbance caused by drain
installation (smear eect), depending on the particular characteristic of soil
and drain to install. Analytical solutions have been studied by many re-
searchers (Barron, 1948; Hansbo, 1979; Van Impe, 1989; Lancellotta, 1995).
However, all of them assume basically constant consolidation parameters in
time.
3
Discussion on undrained shear strength
of soft soils
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
3.1 Preamble
The execution of a staged construction design on soft normally consolidated
soil implies the assessment of the stability in undrained conditions. Therefore,
the initial in-situ undrained shear strength (c
u
) of the cohesive foundation soil
and the subsequent changes of c
u
during construction have to be estimated. In
this chapter, some issues linked to the shear strength of soils and progressive
gain with consolidation are brought to discussion.
3.2 Shear behavior of normally consolidated clay
The shear strength of normally consolidated clays in both drained and
undrained conditions is traditionally represented by Mohr-Coulomb envelopes,
characterized by a friction angle

and an eective cohesion c

which is usually
negligible. Data from a large number of clays from various origins show that
the friction angle is a function of the plasticity of the clay and is independent
of the rate of strain (Fig. 3.1).
Normally consolidated clays behave as plastic materials, consequently,
their stress strain behavior is governed by plastic ow rules; therefore, the de-
nition and use of an elastic modulus of deformation is not justied (Tavenas,
1987). Yet, the stress-strain behavior of normally consolidated clay is often
expressed in terms of undrained modulus (e.g. E
u
).
According to the soil yielding concept, a normally consolidated sample
sheared under undrained conditions follows an eective stress path along the
yielding surface. If the sample is one dimensionally consolidated, as occurs in
nature, the yielding surface is as that shown in gure 3.2.
Roscoe et al. (1958) have shown, based on isotropically consolidated re-
molded clay, that the yielding surface may be described as an elliptical curve
centered on the isotropic axis (Cam-Clay model). However tests on natural
12 Discussion on undrained shear strength of soft soils
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Plasticity index (PI), %
F
r
i
c
t
i
o
n

a
n
g
l
e

'
,

d
e
g
r
e
e
s
From Bjerrum & Simons, 1960
Data from:
Bjerrum & Simons, 1960
Leroueil et al., 1983
Cox, 1970
Jamiolkowski et al., 1979
Figure 3.1. Variation of the eective friction of NC clay with the Plasticity Index
(Mesri and Abel-Ghaar, 1993)
q'
p'
Modified
Cam-Clay
Natural clay
CSL
Ko-line
Figure 3.2. Typical yielding surface of natural clay
clay, which may be assumed to be K
0
-consolidated, show a yield surface cen-
tered on the K
0
-line (Fig. 3.2). The shape of the yield surface is inuenced by
K
0
which in turn is determined by

, closely related to the plasticity index.

Figure 3.3 shows the eective stress path of a set of three samples of clay,
K
0
-consolidated at dierent vertical stresses, subjected to undrained com-
pression. In the gure, p

= (

1
+

3
)/2 and q

= (

3
)/2. By geometrical
means we can identify the initial vertical eective stress on each sample and at
the same time we can identify the undrained strength obtained after shearing.
It can be observed that for a normally consolidated clay, the ratio c
u
/

v
is a
constant, i.e.:
S =
_
c
u

v
_
(1)
=
_
c
u

v
_
(2)
=
_
c
u

v
_
(3)
(3.1)
This important founding led to a better understanding of the clay behavior
subjected to undrained shear. Moreover, it has led to one of the most impor-
tant concepts in soil mechanics, normalized behavior, which is described in
the next section.
Normalized behavior 13
cu
1
v
'
1
'
2
'
3
v v
p'
q'
cu
2
cu
3
Ko-line
CSL
Figure 3.3. Eective stress path in undrained triaxial compression on NC samples
0 5 10 15 20 25
1,0
0,8
0,6
0,4
0,2
0,0
Shear strain, %
(

'
1
-

'
3
)
/

'
c
OCR=8
OCR=4
OCR=2
OCR=1
Figure 3.4. Normalised stress-strain behaviour in triaxial testing (from Ladd and
Foot, 1974)
The undrained strength ratio S = c
u
/

v
obtained from triaxial compres-
sion tests does not vary signicantly with the plasticity index and is typically
within a narrow range. A typical value of S is 0.3 (Tavenas, 1987).
3.3 Normalized behavior
The work carried out by Roscoe et al. (1958), Henkel (1960) and Ladd and
Foot (1974), mostly on reconstituted and destructured clays has shown that
undrained tests on samples having the same overconsolidation ratio but dier-
ent maximum past pressures

p
present similar stress-strain and pore pressure-
strain characteristics when normalized with respect to the consolidation stress

c
or the maximum past pressure

p
. Figure 3.4 presents data from Ladd and
Foot (1974).
These observations gave birth to what Ladd and Foot (1974) called later
the Normalized Soil Parameter (NSP) concept. This concept is very powerful
since it applies to all kind of tests: triaxial, plane strain, direct shear, etc.,
14 Discussion on undrained shear strength of soft soils
in undrained as well as in drained conditions. The NSP is one of the most
fundamental concepts in soil mechanics.
In practice, normalized behavior is not as perfect as that shown in gure
3.4. There is usually some divergence in the normalized plots obtained for
dierent consolidation stresses and also some due to heterogeneity in the soil
deposit. Inevitable minor variations in the procedure from one test to another
can also cause divergence.
Anyhow, normalized behavior has been found to apply to a wide range of
cohesive destructured soils. Still, it is recommendable to check if it is valid for
a more specic case.
Quick clays and highly structured clays will not show a good normalized
behaviour because their structure is usually altered when subjected to recon-
solidation beyond the current stress.
3.4 Shear behavior of overconsolidated clay
When dealing with overconsolidated clays it should be observed that during
unloading the soil has lower water content at the same eective stress than a
normally consolidated material. As a consequence the undrained strength ratio
S, previously dened, should increase with the overconsolidation ratio (OCR).
From data collected by Ladd and Foot (1974) and as shown in gure 3.5,
it appears that the undrained shear strength of destructured clays varies with
the overconsolidation according to:
_
c
u

v
_
OC
=
_
c
u

v
_
NC
OCR
m
= S OCR
m
(3.2)
Ladd and Foot (1974) reported m values approximately equal to 0.8. Jami-
olkowski et al. (1985) indicated that the preceding equation could also be a
good approximation for intact natural clays. Equation 3.2 is also supported
by the critical state theory as shown below.
The critical state line can be dened in the space p

: q

and v : p

as
follows (Atkinson and Bransby, 1978):
q

= M p

(3.3)
v = ln p

(3.4)
where, v is the specic volume, is the specic volume at p

=1 kPa and is
the slope of the normal consolidation line. M, , and , dened in gure 3.6,
are soil constants.
Lets consider two specimens A and B as shown in gure 3.6. Specimen A
is NC while specimen B has a OCR=p

A
/p

B
. The failure states on the critical
state line are indicated by points C and D respectively. The undrained strength
Shear behavior of overconsolidated clay 15
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
1 10
OCR
U
n
d
r
a
i
n
e
d

s
t
r
e
n
g
t
h

r
a
t
i
o

(
S
)
TC; S=0.32 m=0.78
DSS; S=0.28 m=0.78
TE; S=0.20 m=0.82
Figure 3.5. Undrained strength ratio (S = c
u
/

v
) from CK
0
U tests (Jamiolkowski
et al., 1985)
ln p'
Normal consolidation line
(slope -)
v
CSL
Swelling line
(slope -)
ln p'A ln p'B
A
C
B
D
v
B
v
A
Figure 3.6. Undrained shear compression of overconsolidated sample (Atkinson
and Bransby, 1978)
(c
u
)
A
of sample A, which fails on the critical state line with an eective mean
normal stress p

C
and specic volume v
A
, is:
(c
u
)
A
= M p

C
= M e
v
A

(3.5)
16 Discussion on undrained shear strength of soft soils
The swelling line AB passes through point A and has the following equa-
tion:
v = v
A
+ ln
p

A
p

(3.6)
Now, if sample B is allowed to swell to a specic volume v
B
, by analogy
with equation 3.5, the undrained strength (c
u
)
B
will be:
(c
u
)
B
= M e
v
B

(3.7)
Equation 3.6 can be substituted in equation 3.7 putting v =v
B
, it gives:
(c
u
)
B
= M e
v
A

ln
p

A
p

B
(3.8)
(c
u
)
B
= M e
v
A

_
p

B
p

A
_

= (c
u
)
A
_
p

B
p

A
_

= (c
u
)
A
OCR

(3.9)
If we divide the previous equation by p

B
, and keeping in mind that
OCR=p

A
/p

B
, we obtain:
(c
u
)
B
p

B
=
(c
u
)
A
p

B
OCR

=
(c
u
)
A
p

A
OCR
1

(3.10)
(c
u
)
B
p

B
=
(c
u
)
A
p

A
OCR

with =

(3.11)
We can see that equation 3.11 shows a close relationship with equation 3.2
since we are relating normalized parameters of OC to NC clay. The plastic
volumetric strain ratio , as named by Schoeld and Wroth (1968), is de-
termined from the slopes of the compression and swelling lines. Its value is
limited to be between 0 and 1, and is typically about 0.8.
3.5 Laboratory testing techniques
3.5.1 Overview
The main laboratory testing technique developed for the use with the NSP
concept involves consolidation to stresses in excess of those in-situ in order to
overcome sample disturbance eects and to control the OCR.
The simplied eect of sample disturbance is shown in the idealized void
ratio vs. log eective stress plot illustrated in gure 3.7. The virgin com-
pression curve is typically a unique relationship for a specic clay, time of
consolidation and type of consolidation stress system. If a sample becomes
overconsolidated, its eective stress is reduced and it swells, typically, fol-
lowing a relationship such as line a in the gure. With reconsolidation, the
relationship will follow line b back to the virgin compression line.
Laboratory testing techniques 17
Virgin compression
line
Line b
Line a
1 2
3
4
Typical relationship for
disturbed sample during
reconsolidation
Vertical effective stress (log scale)
V
o
i
d

r
a
t
i
o
Figure 3.7. Idealised plot showing the eect of sample disturbance (Ladd and Foot,
1974)
Since the changes in void ratio associated with soil swelling are much
smaller than those associated with virgin compression, overconsolidated soils
always plot below the virgin compression line.
An undisturbed sample will typically suer a decrease in eective
stresses during sampling even though the water content may be kept vir-
tually constant. Thus, an in-situ normally consolidated sample at point 1 (in
Fig. 3.7) might plot at point 2 after sampling and be similar to an overconsoli-
dated sample. With reconsolidation it will follow some path back to the virgin
compression line (dashed line in Fig. 3.7). It follows that a test performed at
conditions corresponding to any point on this line prior to its reaching the
virgin compression line (e.g. a sample reconsolidated to the in-situ stress) has
an uncertain OCR.
On the other hand, a sample that has been consolidated back to the virgin
compression line has a clearly known OCR=1. This sample will give NSP
values which, assuming the concept holds for the soil, are equally applicable to
all normally consolidated samples. If NSP values for overconsolidated samples
are required, these can be obtained at known OCR values by consolidating the
samples back to the virgin compression line and then reducing the eective
stress to the required OCR. This is shown in gure 3.7 as consolidation from
point 2 to point 3, followed by unloading to point 4 to give a sample of known
OCR.
Thus the testing procedure to yield NSP values requires that the sam-
ples be consolidated back to the virgin compression line before testing. Con-
solidation to stress levels greater than 1.5 to 2 times the in-situ stress are
recommended by Ladd and Foot (1974). Moreover, they provide a procedure:
1. Consolidate samples to approximately 1.5 times, 2.5 times and 4 times the
in-situ vertical stress and measure c
u
/

v
. A clay exhibiting normalized
18 Discussion on undrained shear strength of soft soils
behavior will yield a constant value of c
u
/

v
, at least at the two higher
stresses. If c
u
/

v
varies consistently with stress, the NSP concept does
not apply to the clay.
2. To obtain c
u
/

v
vs. OCR, use the minimum value of

v
giving normalized
behavior as the laboratory preconsolidation pressure and perform tests at
OCR values of 2 0.5, 4 1 and 6 2. Compare the results to those
plotted in gure 3.5 to check their reliability. The data point should form
a smooth concave upward curve
It follows that use of this method requires a knowledge of the in-situ
stresses and preconsolidation pressure values; high quality oedometer tests
are essential.
3.5.2 SHANSEP
Stress History And Normalized Soil Engineering Properties (SHANSEP) is
the basis of the technique. It consists of evaluating the stress history of the
clay deposits by evaluating proles of vertical eective stress (

v
) and precon-
solidation pressure (

p
) to determine OCR proles through the deposit and
then applying the appropriate normalized values to give the representation of
strength properties for design.
The basic steps are as follows (Ladd and Foot, 1974; Jamiolkowski et al.,
1985):
1. Examine and subdivide the soil prole into component deposits on the
basis of boring logs, visual classication, etc.
2. Obtain good undisturbed samples and investigate the stress history of
the soil prole using a program of total unit weight, pore pressure and
vertical eective stress measurements. Check whether or not normalized
behavior applies to the soil studied.
3. Perform a series of CK
0
U (one-dimensionally consolidated) shear tests
on specimens consolidated beyond the in-situ preconsolidation pressure
(to

v
greater than 2 times

p
) to measure the behavior of normally
consolidated clay and also on specimens rebounded to dierent OCRs to
measure overconsolidated behavior.
4. Express the results in terms of normalized soil parameters (NSP) and
establish NSP vs. OCR relationships, e.g. c
u
/

v
vs. OCR.
The resulting SHANSEP design strength parameters are expressed in
terms of:
c
u

v
= S OCR
m
(3.12)
in which S and m vary with the in-situ mode of failure as will be seen in the
next sections.
Ladd (1991) made some recommendations for the assessment of S and m.
He stated that CL and CH clays (in the unied system of soil classication)
Laboratory testing techniques 19
tend to have lower, less scattered undrained strength ratios than soils plotting
below the A-line. Moreover, he concluded the following range of typical values,
based on experience:
Sensitive marine clay (PI <30, IL>1): c
u
/

p
=0.20, with a nominal stan-
dard deviation (SD) of 0.015 and m=1.
Homogeneous CL and CH sedimentary clay of low to moderate sensitivity
(PI =20%80%): S =0.20 +0.05PI, or simply S =0.22. Moreover, m=
0.88(1 C
s
/C
c
) 0.06SD, or simply m=0.8.
Northeastern varved clay: S = 0.16 and m = 0.75.
Sedimentary deposits of silts and organic soils (Atterberg limits plot be-
low the A-line, but excluding peats) and clay with shells: S =0.25 with a
nominal SD = 0.05. Moreover, m = 0.88(1 C
s
/C
c
) 0.06SD, or simply
m = 0.8.
Jamiolkowski et al. (1985) stated that there exist evidence to support that
for low OCR inorganic clay of low to moderate plasticity, S falls in a very
narrow band of 0.23 0.04.
3.5.3 Discussion
The SHANSEP technique has been used in a number of projects (e.g. Kout-
softas, 1981; Koutsoftas and Ladd, 1985; Gaberc, 1994; Lechowicz, 1994).
Koutsoftas and Ladd (1985) studied the suitability of the method and
compared it to the conventional practice at that time. They carried out en-
gineering studies for the design of the foundation of an oshore structure on
marine clay.
The conventional practice of testing included Unconsolidated Undrained
triaxial compression tests to obtain the initial c
u
prole and isotropically
consolidated-undrained triaxial compression tests to predict strength increase
with consolidation. The authors concluded the following:
Predictions of the initial c
u
prole based on Unconsolidated Undrained
compression data can easily be in error by 30%.
Predictions of the rate of strength gain with consolidation, based on
isotropically consolidated triaxial tests, will usually be unsafe by 30 10%.
When the project does not warrant development of anisotropic strength
parameters, design values of c
u
can be obtained from one-dimensionally
consolidated direct simple shear tests or from equation 3.2 assuming S =
0.22 0.03 and m = 0.8. Both being reasonable estimates.
The SHANSEP technique provides a good or slightly conservative indica-
tion of stability, whereas, the conventional practice was seen to be erratic in
its prediction of stability and frequently would lead to inadequate designs
of either an unsafe or overconservative nature.
Nevertheless, there also are dierent opinions about the reliability of the
method especially when structured soils are tested. Tavenas et al. (1987) state
20 Discussion on undrained shear strength of soft soils
that the structure of the clay is destroyed when the yielding surface is shifted,
that is, when the eective stress exceeds the preconsolidation pressure.
The SHANSEP technique, in an attempt to overcome the eect of sampling
disturbance, consolidates the samples to stress levels higher (2.5 to 4 times)
than those in-situ. Therefore, compared with actual behavior, the SHANSEP
approach would thus give a smaller stiness and shear strength. However,
when dealing with young soil, where structure is not well developed, the
approach should be more appropriate.
Another limitation of the method is that it can only be applied to fairly
regular deposits for which a well dened stress-history can be obtained. Ladd
and Foot (1974) recommend not to employ the technique when highly hetero-
geneous deposits are encountered; in that case, the study should be supported
by additional eld testing.
3.6 Undrained strength anisotropy
The behavior of soil deposits can be anisotropic because of either structural
anisotropy or stress induced anisotropy. Natural, sedimentary, clays are usu-
ally structurally anisotropic due to the manner of soil deposition during the
formation process; particles tend to become oriented in the horizontal direc-
tion during one dimensional deposition. However, macroscopic variations in
fabric may also produce inherent anisotropy (e.g. sti ssured clay, varved
clay, etc.).
Early research on this eld attempted to evaluate structural anisotropy
by testing samples cut at dierent orientations () to the vertical. Figure 3.8
illustrates the anisotropic nature of the undrained strength measured from
Unconsolidated Undrained triaxial tests on high quality samples. Since all
specimens were sheared in a conventional triaxial apparatus along similar
eective stress paths, the reduction in strength compared to vertical loading
can be attributed to a preferred particle orientation (structural anisotropy
alone).
Soils can also exhibit a stress induced anisotropy whenever K
0
is not equal
to unity (K
0
-consolidation). A structurally isotropic material may show stress
induced anisotropy. It arises from a dierence in the normal stress acting in
various directions as a result of either special loading conditions or bound-
ary conditions. With reference to triaxial testing, we can state that isotropic
consolidation produces stiening in all directions, while, anisotropic consoli-
dation produces stiening in a preferred direction. Consequently, the soil will
respond dierently to loads applied in distinct directions, i.e. compression or
extension.
In practice one deals with both anisotropy components simultaneously.
The practical signicance of anisotropy on staged construction methods is
illustrated in gure 3.9 by considering a long embankment constructed on
soft clay.
Undrained strength anisotropy 21
0 30 60 90
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Angle , (degrees)
c
u
(

)
/
c
u

(
V
)
Normally consolidated San Francisco Bay
Mud (Duncan & Seed, 1966)
Heavily overconsolidated London clay
(Ward et al., 1965)
Overconsolidated Conn.
valleyvarved clay (MIT)
.

Figure 3.8. Undrained strength of three clays measured by UU triaxial compression

(Ladd et al., 1977)
Clay
Embankment
Compression
tests (TC)
Direct simple shear
tests (DSS)
Extension
tests (TE)
Figure 3.9. Relevance of shear tests to shear strength in the eld
Numerical studies also show that foundation clays invariably experience
dierent stress conditions along the potential failure surface (Jardine and
Menkiti, 1999) due to principal axes rotation. The stress conditions might
be represented by triaxial compression (TC), direct simple shear (DSS) and
triaxial extension (TE) at particular points.
The assumption that c
u
is unique function of water content has been dis-
proved by the measurement of signicant strength anisotropy in clay deposits.
For homogeneous non-layered clays sheared under the conditions such as those
in gure 3.9, it is typically found that c
u
from compression tests (TC) is greater
than the value from direct simple shear tests (DSS), which in turn is greater
than the value from extension tests (TE).
Figure 3.10 summarizes values of c
u
/

v
measured in K
0
-consolidated spec-
imens in undrained triaxial compression (TC), extension (TE), and in direct
simple shear (DSS) for dierent clays.
22 Discussion on undrained shear strength of soft soils
0 10 20 30 40 50 60 70 80 90
0.1
0.2
0.3
0.4
0
Triaxial compression (TC): c
u
=q
f
Direct simple shear (DSS): c
u
=
h max
Triaxial extension (TE): c
u
=q
f
Plasticity index, %
(
c
u

/

'
v
)
N
C
TC
DSS
T
E
Figure 3.10. Undrained strength anisotropy from CK
0
U tests on NC clay (Jami-
olkowski et al., 1985)
Figure 3.11. Anisotropy of c
u
/

v
for dierent soil at OCR=1 (Jardine and
Menkiti 1999)
Another means to study anisotropy is through the Hollow Cylinder Appa-
ratus (HCA) which has shown to be an excellent tool of anisotropy investiga-
tion as it allows the control of the principal stress orientation.
Jardine and Menkiti (1999) reported results of their experiences with the
HCA on three silty soils (denoted by HRS, HK, KSS). Figure 3.11 reveals
that the reduction of the undrained strength ratio (S) with the orientation of
principal stresses ( = 0

for compression and = 90

for extension) may be

severe in some soils.
Final recommendations 23
From gures 3.10 and 3.11 we can conclude the following:
Less plastic, and often more sensitive, clays tend to have higher strength
anisotropy than more plastic clays.
The use of an undrained strength ratio estimated from compression tests
(CK
0
U) in stability calculation will yield unsafe results for clays of low to
moderate plasticity index and OCR.
It has been suggested (Jamiolkowski et al., 1985) that most stability analy-
sis should consider strength anisotropy; at least the projects concerning staged
construction or unusual loading. Unfortunately, anisotropy cannot be easily
characterized and implemented in a model for use in routine work. Still, the
assessment of TC and TE tests may help at selecting design parameters even
when a simple model (e.g. isotropic) is employed.
3.7 Final recommendations
This section recommends procedures to obtain undrained strength ratio vs.
OCR relationships required to calculate the initial c
u
prole and subsequent
increases due to consolidation. The recommendations are divided into three
levels of sophistication, as pointed out by Ladd (1991), depending on the
degree of renement required:
Level A: For nal design of all major projects and for sites where the foun-
dation soils exhibit signicant undrained stress-strain-strength anisotropy
or contain unusual features (ssuring, varved, highly organic, etc.) and for
projects requiring accurate predictions.
Level B: For preliminary design and for nal design of less important
projects involving ordinary soils with low to moderate anisotropy.
Level C: For preliminary feasibility studies and to check the reasonable-
ness of initial strengths inferred from in-situ and laboratory conventional
testing programs.
Levels A and B require laboratory CU testing to provide anisotropic and
isotropic (average) input strengths, respectively, whereas level C relies on em-
pirical correlations. Table 3.1 summarizes the testing programs recommended
by Ladd (1991) from his experiences.
The CK
0
U test for Level A use either the SHANSEP technique or simply
Reconsolidation to the in-situ stress, depending on the soil type (e.g. highly
structured), in-situ OCR and sample quality. The c
u
/

v
vs. OCR predictions,
in the form of equation 3.2, should exactly simulate the in-situ response for
the rst stage of construction, but they may involve errors on the safe side
when used to compute strength increases during consolidation.
For Level B programs, Ladd (1991) recommends the use of either CK
0
U
direct simple shear or CK
0
U triaxial compression and extension to estimate
24 Discussion on undrained shear strength of soft soils
Table 3.1. Recommended laboratory testing program (Ladd, 1991)
Level A Level B Level C
CK
0
U tests with CK
0
U tests with either: Uses empirical
dierent modes of failure: Direct simple shear (DSS) correlations
Triaxial compression (TC) or rather than testing.
Direct simple shear (DSS) Triaxial compression (TC) See section 4.5
Triaxial extension (TE) and Triaxial extension (TC) for typical values.
in order to estimate avrg.
strength
a reasonable average value of shear strength along the potential failure sur-
face of a slope. Moreover, he states that isotropic strength proles suce for
the assessment of stability. Level B should not rely on tests performed on
isotropically consolidated specimens. Level C should only rely on empirical
correlations.
Levels A, B and C require a careful assessment of the stress history of the
foundation soil. This fact, plus the observation that c
u
/

v
vs. OCR for most
homogeneous soils falls within a fairly narrow range, means that consolidation
testing usually represents the single most important experimental component
for the design of staged construction projects.
4
Discussion on slope stability evaluation
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
4.1 Preamble
The objective of the present chapter is to review slope stability methods and
related issues. The methods have been classied here into two categories: limit
equilibrium methods and strength reduction methods.
The assessment of the stability of slopes remains a challenging task of
geotechnical engineering. However, many aspects have been thoroughly stud-
ied over the last decades and today the methods of analysis are able to tackle
complex problems.
Slopes, natural or man-made, are observed to collapse in dierent ways.
Figure 4.1 summarizes some of the most common patterns of soil slope failures.
Rocks and soft rocks slopes show dierent patterns out of the scope of this
book.
The two major types of slides are rotational slides and translational slides
(Fig. 4.1). Rotational slide are those in which the surface of sliding is curved
concavely upward and the slide movement is roughly rotational about an axis
parallel to the ground surface. On the other hand, a translational slide is one
in which a soil mass moves along a roughly planar surface with little rotation.
Such planar movement could be the result of the presence of a weak layer or
an interface of dierent soil types.
Moreover, earthows and creep are patterns observed especially in soft ne
grained soils. Earthows have a characteristic shape. They occur for example
when the slope material liquees and runs out forming a depression at the head
and a mound at the toe. On the other hand, creep manifests as a imperceptibly
slow, steady downward of the slope caused by for example, the environment
action, presence of existing sliding surfaces and vicinity of stress state to
failure.
From these 4 slope failure types, the rotational slide and translational
slide were explicitly studied in classic soil mechanics by means of limit states
methods. A short review is given in the next sections. Moreover, the stability
analysis of quick clay masses is evaluated in detail.
26 Discussion on slope stability evaluation
Rotational Landslide Translational landslide
Earthflow Creep
Figure 4.1. Common patterns of soil slope failure (source USGS)
4.2 Causes of slope instability
When facing a design task it is important to understand the causes of insta-
bility of a slope to anticipate the changes in the properties of the soil that may
occur over time, loading conditions, seepage conditions to which the slope will
be subjected, etc.
As stated by Duncan and Wright (2005), when discussing the causes of
slope failure it is useful to start from the very fundamental premise that
the shear strength of the soil must be greater than the shear stress required
at equilibrium. Consequently, the most fundamental cause of instability is
that for some reason, the shear strength of the soil is less than the shear
strength required for equilibrium and such condition can be reached in two
ways:
Through a decrease of shear strength in the soil
Through an increase of the shear stress required for equilibrium
Causes of slope instability 27
Figure 4.2. Cracking as an eect of environment action and its implication in slope
stability
Figure 4.3. Results from fresh water leaching tests on clay specimens from Norway
(Bjerrum, 1967)
Reasons for a decrease in the shear strength of the soil are for example
an increase of pore water pressures (e.g. due to rainy seasons), cracking (i.e.
due to the action of the environment and tension stresses, Fig. 4.2), swelling,
leaching (Fig. 4.3), strain softening behavior, cyclic loading (e.g. leading to
liquefaction). Figure 4.3 illustrates a striking example of soil behavior. These
are results of leaching tests after Bjerrum (1967); they showed at that time
how the structure of a marine deposited soil could signicantly collapse when
it is leached with fresh water creating sensitive clays. This particular example
should always remind the engineer to stay alert and pay attention to the
various ways of soil behavior. Special considerations on the analysis of quick
clays are given in section 4.5.
Reasons for an increase of the shear stress required for equilibrium are for
example an extra loading acting on the slope, water accumulation in cracks
(Fig. 4.2), increase of the unit weight of the soil (e.g. due to wetting), exca-
vation works at the toe of the slope, drop in water level at the site (e.g. due
to water pumping), earthquake or other type of dynamic loading, etc.
28 Discussion on slope stability evaluation
In reality, slopes will fail usually because of a combination of some of the
reasons cited above.
4.3 Stability conditions for analysis
The rst requirement to perform slope stability analysis is to formulate cor-
rectly the problem. Selecting appropriate conditions for analysis of slopes
requires considerations of the shear strength of soils under drained and
undrained conditions, or under drainage conditions that will occur in the
eld.
The general principles involved in selecting analysis conditions and shear
strengths are summarized in table 4.1.
When an embankment is constructed on a clay foundation, the embank-
ment load causes the pore water pressure in the clay to increase. After a period
of time, such increment will gradually dissipate and eventually the pore wa-
ter pressures will return to the initial steady value. As the excess pore water
pressure dissipates, the eective stresses in the foundation soil increase, the
strength of the clay increases and as a result the factor of safety increases
too. Figure 4.4 illustrates these relationships and out of it one may conclude
that the most critical condition occurs at the end of construction (undrained).
Then, it is only necessary to analyze the end-of-construction condition.
Table 4.1. Shear strength for stability analysis (Duncan, 1996)
Condition
End of construction Staged construction Long term
Procedure and Eective stress anal. Eective stress anal. Eective stress anal.
strength for with c

and

with c

and

with c

and

sand
Procedure and Total stress anal. Total stress anal. Eective stress anal.
strength for with c
u
from approp. with c

and

clay consolidation anal.

When a slope in clay is created by excavation, the pore pressures in the clay
decrease in response to the removal of the excavated material. Over time, the
negative excess pore water pressure dissipate and the pore pressures eventually
return to the initial steady value. As the pore water pressure increases, the
eective stress in the decreases and the factor of safety decreases with time
as illustrated in gure 4.4. Out of these relationships it can be concluded that
the long-term (drained) condition is more critical than the end-of-construction
condition.
Drained conditions are analyzed in terms of eective stresses using val-
ues of c

and

determined from drained tests, or from undrained tests with

Stability conditions for analysis 29
Figure 4.4. Stability conditions for analysis
pore water pressure measurements. When dealing with clay, drained triaxial
tests are frequently impractical because the required time is very long, there-
fore, undrained triaxial tests with pore pressure measurements are the most
suitable. Values of

for natural deposits of cohesionless soils are usually esti-

mated using correlations from eld tests results (i.e. CPT, SPT, DMT, PMT)
given the current diculties of testing high quality undisturbed sand samples.
Undrained conditions are analyzed in terms of total stress in order to
avoid having to rely on estimated values of pore water pressure for undrained
loading conditions. Undrained shear strength of soils is usually correlated from
eld tests or laboratory tests. For staged construction analysis, the undrained
strength is furthermore assessed through consolidation analysis in combination
with for example triaxial CU testing.
In cases where it is not clear whether the short-term or long-term condition
will be more critical, both should be analyzed, to ensure that the slope will
have an adequate stability under any condition.
Another important topic is that of the selection of an suitable factor of
safety for design. The main considerations to take into account are the degree
of uncertainty in evaluating conditions and shear strengths for analysis and
the possible consequences of failure. Typical minimum acceptable values of
30 Discussion on slope stability evaluation
factor of safety are about 1.3 for the categories end-of-construction and staged-
construction and 1.5 for long-term conditions.
4.4 Stability analysis procedures
The universal availability of computers and a much improved understanding
of the mechanics of slope stability analysis have brought about considerable
changes in the computational aspects of slope stability analysis in the last
years.
Figure 4.5. Limit equilibrium approaches
4.4.1 Limit equilibrium methods
The calculation of global stability is commonly expressed in terms of the Fac-
tor of Safety computed by means of Limit Equilibrium Methods. The principles
underlying these methods are as follows:
A slip mechanism is postulated (i.e. a potential failure surface is outlined).
The shearing stress to equilibrate the assumed slip mechanism is calculated
by means of statics.
The calculated shearing stress required for equilibrium is compared to the
available shear strength in terms of the factor of safety.
The mechanism with the lowest factor of safety is found by iteration.
Stability analysis procedures 31
Figure 4.6. Free-body equilibrium of slices: a) ordinary method of slices, b) sim-
plied Bishop procedure, c) more rened methods
In general, two approaches to study the statics of limit equilibrium meth-
ods can be identied: the single free-body approach and the method of slices.
The single free-body approach considers one single free body that slides down-
ward. The innite slope and the Swedish circle are examples of such analysis
(Fig. 4.5) although their applicability is limited in practice. On the other
hand, the method of slices includes such methods as the Ordinary Method of
Slices, Simplied Bishop procedure and others. Unlike single-free body anal-
ysis, it subdivides the soil mass prone to sliding in slices and considers the
equilibrium of each of them. Figure 4.6 illustrates the forces acting on the
slices according to some methods. The simplest methods do not consider any
interslice forces while other more elaborated do.
Some of the method of slices proposed in literature and their satised
equilibrium conditions are summarized in table 4.2.
4.4.2 Strength reduction methods
More recently, the following denition of factor of safety has gained a lot of
acceptance: The Factor of Safety is that factor by which the shear strength pa-
rameters may be reduced in order to bring the slope into a state of failure. The
preceding denition has given rise to a new technique, the so called Strength
Reduction Methods (SRM). Such denition has been easily implemented
on nite element and nite dierence computer programs (i.e. PLAXIS,
FLAC).
The principle behind the shear strength reduction technique in nite el-
ement slope stability analysis is to reduce c

and

or c
u
by a factor until
failure occurs. Consequently, the overall safety factor and the corresponding
potential failure surface can be obtained simultaneously. Finite element slope
32 Discussion on slope stability evaluation
Table 4.2. Characteristics of equilibrium methods of slope stability analysis (Dun-
can, 1992)
Method Characteristics
Slope stability charts Fast and accurate enough for many purposes
Ordinary method slices For circular slip surfaces
Satises moment equilibrium only
Bishops modied method For circular slip surfaces
Satises moment equilibrium
Satises vertical forces equilibrium only
Force equilibrium method For any shape of slip surface
Does not satisfy moment equilibrium
Satises vert. and horiz. force equilibrium
Morgenstern and Price For any shape of slip surface
Satises all equilibrium conditions
Spencers method For any shape of slip surface
Satises all equilibrium conditions
failure prediction by the shear strength technique is performed by using two
reduced shear strength parameters, namely:
c
R
=
c
R
(4.1)
tan
R
=
tan
R
(4.2)
where R is called the shear strength reduction factor.
A starting value of 1 (no reduction of strength) is usually given to R. As the
calculation proceeds, R is increased and the shear strain and displacements
are evaluated for each step until failure is reached. The shear strength factor
(R) at failure is called critical strength reduction factor (R
f
) and corresponds
to the overall safety factor of the slope.
The output of the calculation is expressed as a graph illustrating the
progress of R. Figure 4.7 shows the development of the strength reduction
factor (R) versus the displacement of a control node within a nite element
grid modeling a slope stability problem (Fig. 4.8).
Note that it is only meaningful to refer to a factor of safety when a steady
state solution is obtained. Intermediate values of R do not have any physical
meaning and are only used for numerical purposes. The displacements ob-
tained during the calculations also do not have any physical relevance. So, by
looking at the progress of calculation for a particular slope (Figs 4.7 and 4.8)
it can be deduced that a steady state solution has been clearly obtained at the
end of the calculation as indicated by the at slope of the curve. The factor
of safety is therefore (for this example) determined as 1.6 approximately.
Although the magnitude of total displacements does not have physical
meaning, the displacement pattern calculated for the last step (incremental
Stability analysis procedures 33
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
0 1 2 3 4
Displacement, m
S
t
r
e
n
g
t
h

r
e
d
u
c
t
i
o
n

f
a
c
t
o
r
,

R
Figure 4.7. Strength reduction factor (R) versus displacement.
Figure 4.8. Shadings of incremental displacements delineating the failure mecha-
nism (PLAXIS output)
displacements) provides an idea of the failure mechanism developing in the
slope. Note that the failure mechanism, as shown in gure 4.8 is not fully
circular.
The shear strength reduction technique has a number of advantages over
the limit equilibrium method (Matsui and Ka-Ching, 1992). Probably the
most remarkable advantage is the automatic failure surface determination.
The application of this technique has been limited in the past due to the long
computer run times required. But with the increasing speed of computations,
the use of the technique is also increasing.
In order to investigate the results of the strength reduction method, fac-
tors of safety obtained with PLAXIS have been compared to factor of safety
estimated by the limit equilibrium method. The next section describes this
comparative analysis in more detail.
4.4.3 Limit equilibrium versus strength reduction methods
A number of simulations were performed for a wide range of parameters and
for dierent embankment geometries with a variety of slopes ranging from
15

to 90

. The results were back calculated to obtain coecients or stability

factors from methods generally accepted such as the Taylor method (Taylor,
34 Discussion on slope stability evaluation
1937), which is a total stress analysis, and, the Bishop-Morgenstern method
(Bishop and Morgenstern, 1960), an eective stress analysis.
[Link] Taylors method
Taylor (1937) dened the stability factor N
0
, as shown by equation 4.3, to
prepare a chart in order to determine the stability of slopes in a homogeneous
deposit of soil underlain by a much stier strata.
N
0
= FS
H
c
u
(4.3)
Equation 4.3 was used to estimate the Taylors stability factor (N
0
) from
simulations performed in PLAXIS given the unit weight (), the embank-
ment slope (H), the undrained strength (c
u
) and the factor of safety (FS)
obtained.
The back calculated stability factor (N
0
) was then plotted against the
slope inclination. Figure 4.9 shows the stability chart as proposed by Taylor
(1937) and the back calculations from PLAXIS. One can see that the approach
is close and the trends are similar.
2
3
4
5
6
7
8
9
10
0 10 20 30 40 50 60 70 80 90
Slope angle []
S
t
a
b
i
l
i
t
y

f
a
c
t
o
r
Strength reduction,
PLAXIS
Limit equilibrium,
Taylor (1937)
d=0
d=1
d=
d=D/H
Figure 4.9. Total stress analysis; stability factor versus slope angle
The points calculated through the strength reduction method are within
a few percent of the limit equilibrium solution. Nonetheless, one must bear
in mind that Taylors solution refers exclusively to circular surfaces while the
strength reduction method does not impose any restriction to the geometry
of the failure mechanism. Therefore, the slight dierences might be reecting
the limitations of Taylors method instead of the incorrectness of the strength
reduction method.
Failure mechanisms for highly sensitive clays 35
Values for m (d=oo; c'/ H = 0)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
2:1 3:1 4:1 5:1
Slope
m
PLAXIS
Bishop &
Morgenstern(1960)
Ph i=40
Ph i=30
Ph i=20
Figure 4.10. Eective stress analysis; stability coecient m for c/H =0 and
d =
[Link] Bishop-Morgenstern method
Bishop and Morgenstern (1960) prepared a number of charts for homogeneous
soil slopes with simple geometry using Bishops simplied method of slices.
They expressed the factor of safety as follows:
FS = mnr
u
(4.4)
where m and n are stability coecients that depend on the drained friction
angle, the drained cohesion (if exists) and the geometry of the slope. The
procedure also assumes circular potential failure surfaces.
In view of the amount of charts for the method, the analysis carried out
here has been restricted. Only 2 charts were employed here to compare the
results; one of them with c

= 0 and the other with c

= 0.05H. Moreover,
absence of pore water pressure was adopted (r
u
= 0). In that way, the factor
of safety calculated with PLAXIS directly equals to m.
Figures 4.10 and 4.11 show the charts for estimation of the stability coe-
cient m proposed by Bishop and Morgenstern (1960) and the back calculated
points from PLAXIS output. One can see that the agreement is much closer.
Finally, it can be concluded that Strength reduction factors of safety were
within a few percent of the limit equilibrium solution and that in general a
close agreement was observed. The fact that the limit equilibrium methods
employed here (e.g. Taylor, 1937; Bishop-Morgenstern, 1960) assume circular
surfaces of potential failure may be causing such small deviations.
4.5 Failure mechanisms for highly sensitive clays
Van Impe and De Beer (1984)
Highly sensitive clays or quick clays do relate to clays that when remolded
loose their structure completely and then their shear strength (after remold-
ing) is reduced almost to zero.
36 Discussion on slope stability evaluation
Values for m (d=0; c'/ H = 0.05)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
2:1 3:1 4:1
Slope
m
PLAXIS
Bishop &
Morgenstern (1960)
Ph i=40
Ph i=30
Ph i=20
Figure 4.11. Eective stress analysis; stability coecient m for c

/H = 0.05 and
d =1
Sensitivity is the ratio between the undisturbed and the fully remolded
undrained shear strength. In this framework, quick clay can be dened as a
clay with a sensitivity of 50 or more and a fully remolded shear strength of
less than 0.4 kPa.
Quick clays are found in areas once glaciated during the Pleistocene epoch.
They have mainly been located in northern Russia, Norway, Finland, Sweden,
Canada and Alaska. These areas are all characterized by geological isostatic
uplift which took place after the retreat of the ice. The development of very
high sensitivity is often the result of processes that have taken place after the
deposition of the clay layers such as leaching of salts (remotion of dissolved
salts), changes in the ion composition of the pore water, pH of the pore water,
dispersive action of some organic and inorganic natural compounds, etc.
The slope stability analysis in highly sensitive clays as proposed by Norwe-
gian researchers (Bjerrum, 1973; Gregersen, 1981; Aas, 1981) is mainly based
on the drained shear behavior of the soil in or along the the potential sliding
surfaces. It has been reported in involved Norwegian investigations that im-
mediately after the sliding of a quick clay mass, no excess pore water pressures
are measured in the immediate vicinity of the sliding surface. This is often
considered to be an experimental support for using drained parameters in the
stability analysis.
However, the possibility of the existence of excess pore water pressures
in the more pervious silt seams of the undisturbed quick clay layer cannot
be excluded. Such excess pore water pressures can disappear as soon as the
sliding occurs. They however can be generated simultaneously in the quick
clay itself, due to remolding during sliding.
In the Swedish approach, as described by Bernander (1981), use is made of
the undrained shear strength parameters. It is assumed that the shear plane
coincides with the interface of the metastable quick clay layer and the more
resistant substratum. The mentioned contribution stresses the large inuence
Failure mechanisms for highly sensitive clays 37
of the conguration of that contact plane on the stability conditions. In situ
observations show the sliding does not always occur along the contact plane
with the substratum, but also other shear planes are frequently observed.
Analyzing slidings of quick clays, initiated apparently without any external
cause, Bernander concludes some of these slidings must be related to creep
phenomenae described as proposed by Mitchell et al. (1968). If such creep
phenomenae however only should be a consequence of the shape of the contact
plane itself, they should have started from the very beginning of the formation
of the actual geometry of the soil surface and of the involved layers. Such creep
phenomenae consequently exist for several decades and deformation velocity
therefore should be extremely small.
Tests results of Bernander (Fig. 4.12) indicate, for a clay at a given over-
consolidation ratio, that the ratio of the residual shear strength
r
at large
deformation () to the maximum shear stress (
max
) tends toward unity when
the deformation speed (v) decreases signicantly. Therefore, it seems unlikely
the creep phenomenae mentioned before should be able to explain the oc-
currence of a landslide after several centuries of existence of the clay layer
interface in its actual state.

(mm)
v=0.025 mm/s

max
v=0.125 mm/s
v=0.500 mm/s

max
(Deformation speed)
Constant OCR = 1.5
1.0
0.5
5.0 10.0
Figure 4.12. Stress strain curves of clay specimens sheared at dierent strain rates
(after Bernander, 1981)
On the contrary, the activation of creep due to some new external factors
can contribute to such explanation.
Moreover, from literature in general it seems that insucient attention is
paid to the possible existence of excess pore water pressures developed in the
silty seams often detected in such clay masses.
4.5.1 Flake type sliding of quick clay
In order to explain ake type sliding of quick clay masses, the reasoning as
proposed by Aas (1981) is commented rst. For sake of simplicity the case of
38 Discussion on slope stability evaluation
'
'
mob
='
crit
'
3
'
h
u
u='
vo
-'
v
'
v
S
P
O
1
O
2
'
v
vo
'
crit
'
h
2
2
*
2
*
A
B
C
D
H
'
v
'

h
'
h
C
P Q

h(BD)
'



'
'
vo
-u
a) b)
c)
'
' '
1
'
Figure 4.13. Stress state at a horizontal potential sliding surface
horizontal soil surface and horizontal potential sliding plane BD is considered
(see Fig. 4.13a). In the original situation at rest, the eective stresses

v,0
and
K
0

v,0
are assumed in the vicinity of the plane BD. In addition, by applying
an horizontal external force H, shear stresses
h
are introduced.
From the Mohr circle 2, in gure 4.13b, representing the stress state in the
plane BD, one gets:

h
=

3
2
_
2

v,0
u

1
+

3
2
_
2
(4.5)
and

1
+

3
2
=

v
+

h
2
=

v,0
u +K
0

v,0
u
2
(4.6)
or

1
+

3
2
=

v,0
1 +K
0
2
u (4.7)

v,0
u

1
+

3
2
=

v,0
1 K
0
2
(4.8)
Failure mechanisms for highly sensitive clays 39
If the angle of tangent through the origin at the Mohr circle 2 is designated

mob
, the following can be written:

3
2
=

1
+

3
2
sin

mob
(4.9)
or

3
2
=
_

v,0
1 +K
0
2
u
_
sin

mob
(4.10)
and therefore:

h
=

v,0

_
_
1 +K
0
2

u

v,0
_
2
sin
2

mob

_
1 +K
0
2
_
2
(4.11)
The equation 4.11 has a general validity. From it, the variation of excess
pore water pressure with sin

mob
, when

v,0
and
h
are given values, can be
deduced:

sin

mob

_
1 +K
0
2

u

v,0
_
2
sin
2

mob
_
1 +K
0
2
_
2

= 0 (4.12)
Starting with the value of u = 0 at the initial stress conditions, one obtains
from equation 4.12:

v,0
sin

mob
=
1 +K
0
2 sin

mob
(4.13)
On gure 4.14b after Aas, from CU triaxial tests performed on ve dierent
normally consolidated clays, the results are shown of the variation of a pore
water pressure function F(u) versus the values of the mobilized angle

mob
of
internal friction.
The curves with respect to the indicated numbers 4 and 5 for non-sensitive
clays are from a certain level for sin

mob
quite dierent from those with
respect to the sensitive clays (numbers 1, 2 and 3). For the last mentioned,
starting at values sin

mob
0.42 to 0.45, a marked rising of the pore water
pressure, and a corresponding decrease of the shear strength (Fig. 4.14a) is
found.
After reaching the critical

mob
value of the angle of friction, the shear
strength of the quick clay drops drastically (see Fig. 4.14a) because the further
rising up of the mobilized friction angle

cannot compensate the decrease of

the eective stress

caused by the rising of the excess pore water pressure

u. On the contrary, that is not the case for two non-sensitive clays (4 and 5).
So, if for quick clays along whatever plane, the critical value of

is mobi-
lized, any small increase of the shear stress will cause an almost total reduction
of the shear strength.
40 Discussion on slope stability evaluation
sin '
mob
1, 2, 3: Sensitive clays; 4, 5: Normal clays
0.4
-0.1
0.2 0.3 0.4 0.5 0.6 0.7
0.0
0.1
0.2
0.3
sin '
mob
0.5
0.0
0.1
0.2
0.3
0.4
2
1
3
4
5
1
2
3
4
5
a)
b)
sin '
crit
=0.42 to 0.45
s
l
=
(

1
-

3
)
/
2

'
v
o
F
(
u
)
=
u
/

'
v
o
-
(
k
o
-
1
/
2
)
Figure 4.14. Variation of shear strength and pore water pressure function F(u)
with mobilized shear angle
When for kinetic reasons only a continuous sliding along the horizontal
plane BD should be possible, the point representing the eective stress on
that horizontal plane in the Mohr circle (for normal clays at sliding conditions)
should be located on the intrinsic straight line OQ at a friction angle

(Fig.
4.13c). When the values of

and
h
are given, the possibility of sliding along
the horizontal plane needs the development of an excess pore water pressure
u given by:

hBD
=
_

v,0
u
_
tan

(4.14)
In quick clays however the conditions for continuous sliding along the
horizontal plane BD are much less severe. Indeed, when in such clays along
an arbitrary plane (not necessarily a kinematically possible), the critical value

crit
=

mob
<

of the friction angle is mobilized, any very small increase

of the shear stress will cause a drastic decrease of the shear strength and a
subsequent liquefaction.
From the previous considerations it results for quick clays the Mohr circle
2 in gure 4.13b representing a critical stress state. The ratio of the shear
stress
h
to the vertical eective stress

v
on the sliding plane (Fig. 4.13b) is
dened by an apparent angle

, much smaller than

mob
.
4.5.2 Analysis taking pore water pressures into account
Suppose a mass of quick clay with inclined surface at a small angle and
consider a potential sliding plane parallel to the mentioned ground level
Failure mechanisms for highly sensitive clays 41
'

0
b)
M
M'
N
'
*
S

,crit

'
3
Co '
1
'
0
3
2
1

A
B
C
D
h
Gn

N
T

Potential
shear plane
Phreatic level =
surface level
l
Eo
Eo
W
W
W
BD
a)
Figure 4.15. Stress state of a inclined slope with water level at ground surface
(Fig. 4.15a). The phreatic level in each point is thought to correspond with the
ground surface, then a groundwater ow parallel to the surface takes place.
When the slope is innitely extended, each vertical section is a symmetrical
one and the indicated forced W and E
0
are independent of the considered
section.
For the quick clay soil mass ABCD of gure 4.15a, one gets a tangential
force T:
T = G
n
sin (4.15)
or
T =
s
h l sin cos (4.16)
where
s
is the unit weight of the soil.
So a shear stress value
0
is working on BD:

0
=
s
h cos sin (4.17)
The total normal force N on BD is given by:
N =
s
h l cos
2
(4.18)
Taking into account the uplift force of the water:
W
BD
=
w
h l cos
2
(4.19)
42 Discussion on slope stability evaluation
The initial eective stress

0
on the plane BD becomes:

0
= (
s

w
) h cos
2
(4.20)
For the normally consolidated sensitive clay, the equilibrium conditions
may be written as:

0
tan

(4.21)
Rewriting equation 4.21, considering the expressions 4.17 and 4.20, this
equilibrium condition is given by:
tan

s

s
tan

(4.22)
It can be seen from gure 4.15b, the resulting eective stress on BD is
inclined over an angle toward the normal on BD;
tan =

0

0
=

s

w
tan (4.23)
About the value of the angle of internal friction

of the undisturbed
soil skeleton to be considered, extensive research by means of dierent kinds
of laboratory tests was performed (Bjerrum, 1969; Aas, 1981). Out of the
Norwegian test results it seems the

value to be taken into account for

the shear strength of undisturbed Norwegian quick clay skeleton is about

mob
=

25

, that is sin

mob
0.42 (see Fig. 4.14).
If, for example,
s
=18.5 kN/m
3
and
s

w
=8.5 kN/m
3
, in order to fulll
the equilibrium condition, the angle (Fig. 4.15a) of the inclination of the
sliding surface must be limited to (Eq. 4.22):
tan
8.5
18.5
tan

(4.24)
or
tan 0.21 (4.25)
or
12

(4.26)
This means, in the case without considering any pore water pressure de-
velopment, the slope stability could be guaranteed for a value up to (Fig.
4.15b):
tan
SN
SO
= tan

= 0.466 (4.27)
or

25

(4.28)
Thus, if in case of seepage, a surface with a slope of = 25

should be
possible; however in case of the existence of a seepage parallel to the ground
Failure mechanisms for highly sensitive clays 43
surface, the slope angle is reduced to =12

. When, however, the water pres-

sures are artesian, the stability conditions can become much more unfavorable.
Let us now consider a quick clay mass with surface inclined at an angle
satisfying the inequality 4.25 and therefore it is in an equilibrium state.
On the clay mass supplementary disturbing forces are applied such that, on
the most unfavorable elementary plane in the considered point, there exists
an eective stress represented by point M on Mohr circle 2 in gure 4.15b
is located on the intrinsic line with an angle

. In normal clays these most

unfavorable elementary planes do not constitute a kinematic possible sliding
plane, as it is supposed that such a sliding only is possible along a plane BD
parallel to the soil surface (Fig. 4.15a). In normal clays the representative
point of the maximum shear stress on an elementary plane parallel to BD
should be located in point N of Mohr circle 3 in gure 4.15b. In quick clays a
distinction has to be made between slow loading and rapid loading. Mohr circle
3 is representative of a slow loading condition, on the other hand, Mohr circle
2 is representative of a rapid loading reaching the critical stress conditions
in the elementary plane of the point M where liquefaction is induced. It is
consequently sucient (along the sliding plane BD of Fig. 4.15a) to attain
the shear stress SQ in order to have general sliding. The apparent angle

can be evaluated as follows:

0
tan

3
2
_
2

1
+

3
2
_
2
(4.29)
and with

3
2
=

1
+

3
2
sin

(4.30)
one gets:

1
+

3
2
=

0
1 sin
2

_
1
_
1
_
1 sin
2

_ _
1 + tan
2

_
_
(4.31)
Also:

1
=

0
(1 + sin

)
1 sin
2

_
1
_
1
_
1 sin
2

_ _
1 + tan
2

_
_
(4.32)
and

3
=

0
(1 sin

)
1 sin
2

_
1
_
1
_
1 sin
2

_ _
1 + tan
2

_
_
(4.33)
For the value of
crit
=M

M (Fig. 4.15b) the following expression can be

used:

crit
=

3
2
cos

(4.34)
44 Discussion on slope stability evaluation
From several experimental data, mainly based on laboratory tests on Nor-
wegian clays, the NGI deduced:
tan

= 0.2
or

= 11

18

Introducing in equation 4.31 the values of

25

and

11

one can
obtain:

1
+

3
=
2

0
1 0.18
_
1
_
1 (1 0.18)(1 + 0.04)
_
(4.35)
or

1
+

3
= 1.5

0
(4.36)
From equations 4.32 and 4.33 we get:

1
=

0
(1 + 0.42)
1 0.18
_
1
_
1 (1 0.18)(1 + 0.04)
_
(4.37)
or

1
= 1.07

0
(4.38)
For

3
in the same way it is derived:

3
= 0.44

0
(4.39)
The value of
MM

crit
= MM

is obtained from equation 4.34:

MM

crit
= 0.285

0
(4.40)
To this critical shear stress
MM

crit
on the shear plane corresponds, on the
plane BD, a shear stress
,crit
= SQ (Fig. 4.15b)

,crit
= SQ = tan

0
0.2

0
(4.41)
or

,crit
= 0.2 (
s

w
) hcos
2
(4.42)
4.5.3 Mechanism of sliding in quick clay masses
[Link] Disturbing action by disappearance of downward
supporting forces
In gure 4.16a a quick clay mass ABCD initially in equilibrium is consid-
ered. At the downward end at section AB, suddenly all supporting forces are
assumed to disappear. In a quick clay, a sliding will occur when the shear
stresses along BD reach the critical value
,crit
, given by:

,crit
L = E
a
+W +
s
h L cos sin (4.43)
Failure mechanisms for highly sensitive clays 45
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6
Beta ()
L
/
h
a)
b)
Ea
W
Eo=0
W=0

A
B
C
D
L
h

Ground surface =
phreatic level
Figure 4.16. Variation of L/h as a function of
where E
a
is the active earth pressure at sliding on plane DC:
E
a
= K
a
(
s

w
)
h
2
2
= tan
2
_

2
_
(
s

w
)
h
2
2
(4.44)
From equations 4.42 and 4.43 one gets:
0.2(
s

w
) h L cos
2
= K
a
(
s

w
)
h
2
2
+
w
h
2
2
+
s
h L cos sin
(4.45)
or
L =
h(
s

w
)K
a
+
w
2 cos [0.2(
s

w
) cos
s
sin ]
(4.46)
For the case of an extremely slow change of the stress conditions in the
fully saturated quick clay mass (as shown before in equations 4.22, 4.24 and
4.25) the quick clay could remain in metastable equilibrium for values of the
slope angle not higher than
max
= 12

, for the chosen numerical example. At

this extreme value of any external disturbing force will cause liquefaction
of the sensitive clay mass.
On the other hand, in the case of a relatively quick change of the loading,
as for example the sudden disappearance of the downward supporting forces,
46 Discussion on slope stability evaluation
for quick clays the maximum slope angle
crit
is limited to (from equation
4.22):
tan
crit
=

s

s
tan

(4.47)
Substituting

= 11

18

,
s
= 18.5 kN/m
3
and
w
= 10 KN/m
3
we get:

crit
= 5

16

(4.48)
As it is probable that quick clay masses were subjected at some point in
their geological history to rapid loading conditions, one can expect the natural
inclination of the ground level to be not higher than such limiting value of

crit
5

.
Concerning the variation of the length of the sliding mass as a function of
the inclination , one gets out of equation 4.46 and for assumed values of

s
and
w
:
L
h
=
8.5 tan
2
(45

11

18

) + 10
2 cos (0.2 8.5 cos 18.5 sin )
(4.49)
The variation of L/h versus is illustrated by the curve in gure 4.16b.
For values up to about 4

the value of L/h is rather small. All of this is

based on the assumption of a phreatic level coinciding with the soil surface.
For smaller values of the relative length L/h remains small. In such cases
it is probable a succession of retrogressive shell shaped slidings will occur.
When on the contrary the slope angle reaches its critical value
crit
, the
relative length L/h becomes more signicant and a ake type sliding will be
more likely to result.
[Link] Considering excess pore water pressures built up
from the upper end in more pervious seams
In many descriptions of quick clay landslides (Aas, 1981; Broms, 1983;
Gregersen, 1981) it is mentioned that more pervious seams are often found in
the clay mass. For several reasons it can occur that excess pore water pres-
sures at a given moment are developed from the upper region of the slope in
the more pervious seams.
For example, in the very well know case of the Rissa landslide, a ake type
of sliding occurred after successive shell shaped retrogressive slidings. On the
surroundings of the retrogressive slidings high back scarps of several meters
were reported.
At the start of a shell shaped slide, it can be assumed that the total weight
of the collapsing soil is supported by water, so creating an excess pore water
pressure u:
u =
s
h
shell

w
shell (4.50)
Failure mechanisms for highly sensitive clays 47
b)

A
B
C
D
h
G
Potential
shear plane
Phreatic level =
surface level
L
Ea
E
p
W
W
a)
G sin
u

(E
o
)

h
L
u

h
h x
w
= u
Figure 4.17. Quick clay slope stability analysis taking pore pressure built up into
account
This local excess pore water pressure can penetrate into the still remaining
clay mass and alter the equilibrium state. In order to get an idea of the possible
inuence of the excess pore water pressure a very simplied approach can be
made.
Let us suppose, at the upper end of the soil mas ABCD (Fig. 4.17), an
excess pore water pressure u is built up in the more pervious seams. Assuming
a simple triangular distribution of u and from equations 4.42, 4.43 and 4.45
one gets:

,crit
L +E
p
= E
a
+
s
h L cos sin (4.51)
where E
p
is the passive earth pressure at the lower end of the sliding mass.
When the value of u remains suciently below the value of

0
, it is possible
to write
,crit
as (from Eq. 4.42):

,crit
= 0.2
_
(
s

w
) hcos
2

u
2
_
(4.52)
As in the case of quick clays, liquefaction should already have taken place
before a passive earth pressure state can be reached; in equation 4.51 E
p
is
changed in the neutral earth pressure value E
0
. From equations 4.51 and 4.52
it results:
E
a
+
s
hLcos sin E
0
+ 0.2
_
(
s

w
) hcos
2

u
2
_
L (4.53)
As the angle is at most 5

30

, cos is put equal to 1. Consequently:

s
hLsin E
0
E
a
+ 0.2
_
(
s

w
)h
u
2
_
L (4.54)
48 Discussion on slope stability evaluation
0.05
0.15
0.25
0.35
0.45
0.55
0.65
0.75
0.85
2.5 3 3.5 4 4.5 5 5.5
Beta ()
L/h=oo
L/h=50
L/h=30
L/h=10
L/h=2
h/h

crit
Figure 4.18. Variation of h/h as a function of
with
E
0
E
a
=
_
(1 sin

)
1 sin

1 + sin

_
(
s

w
)
h
2
2
(4.55)
In order to prevent the phenomenon of piping in the clay, the following
condition can be written:
u < (
s

w
)h = 0.85h (4.56)
When u =
w
h, equation 4.56 leads to:
h
h
<

s

w
= 0.85 (4.57)
At h/h = 0.85 piping in the clay mass occurs instead of sliding. For all
values of h/h < 0.85, the expression of sin (with assumed values of

,
s
and
w
) from equations 4.54 and 4.55 becomes:
sin
0.729 + 0.2
_
8.5 5
h
h
_
L
h
18.5
L
h
(4.58)
Varying h/h and L/h, the corresponding values of the slope angle at
limit of equilibrium are calculated and given in gure 4.18.
The application of equation 4.58 is limited up to the value of sin sin

crit
. Moreover, for small values of L/h the assumption of a linear decreasing
excess pore water pressure along the potential sliding surface cannot be valid
anymore.
Risk of liquefaction 49
So, in a quick clay mass with a slope angle close to the critical value and
in which more permeable seams are present, an increasing excess pore pressure
at the upward end can easily explain the occurrence of long ake type slidings.
For such large ake type slidings, the required excess pore water pressure at
a given slope angle is smaller than the corresponding value for shorter shell
shaped slidings. Such conclusions are valid when the slope angle is close to its
critical value
crit
.
4.5.4 Conclusions
Analyzing the stability problem of quick clay masses taking into account excess
pore water pressures, the slope angle seems to be limited to < 12

when
assuming that the phreatic level coincides with the ground surface and stress
conditions are changed slowly. In the case of a relatively quick change of
external loading, such slope angle is reduced to < 5

30

.
For smaller slope angles a succession of retrogressive shell shaped slid-
ings can occur. Even with a very simple assumption of the excess pore water
pressure distribution in more permeable seams within the clay mass an ex-
planation can be found for the occurrence of long ake type slidings. They
are more likely to occur when the undisturbed quick clay mass in nearly at
rupture condition ( close to its critical value).
4.6 Risk of liquefaction
Soil liquefaction is a major concern for structures with or on loose sandy soils.
Liquefaction can be caused by earthquakes by inducing a progressive build-
up of excess pore water pressure due to cyclic shear stresses. When the pore
pressure builds up to a level equal to the initial conning stress, soil loses its
strength and large deformation occurs.
To evaluate the potential for soil liquefaction at a particular site, it is im-
portant to determine the soil stratigraphy and the state of the soils. While
much research has been performed over the past decades to advance the tech-
niques for assessing liquefaction potential of soils (Ishihara, 1993), simplied
procedures are still the most widely used methods (e.g. Seed and Idriss, 1971;
Robertson and Wride, 1998).
In the simplied approach, the cyclic stress ratio (CSR) generated at any
depth of the soil deposit due to the earthquake loading can be obtained using
simplied equations such as:
CSR =

av

v0
= 0.65
_
a
max
g
_ _

v0

v0
_
r
d
(4.59)
where
av
is the average cyclic shear stress, a
max
is the maximum horizontal
ground acceleration at the ground surface, g is the acceleration due to gravity,
50 Discussion on slope stability evaluation

v0
and

v0
are the total and eective vertical stresses respectively and r
d
is
a stress reduction factor which is a function of depth. Seed and Idriss (1971)
proposed:
r
d
= 1.0 0.00765 z if z < 9.15m
r
d
= 1.174 0.0267 z if 9.15m < z < 23m
On the other hand, CSR depends on the maximum acceleration at the
ground surface. If CSR exceeds the cyclic resistance ratio (CRR) at any depth
of the soil deposit, soil liquefaction occurs at that depth. Here, a factor of
safety against liquefaction can be dened as FoS = CRR/CSR. In theory, no
liquefaction is expected to occur if FS > 1; and on the other hand, if FS 1,
liquefaction is expected.
Over the past 25 years, numerous studies have been carried out to correlate
the CRR to in situ tests such as the standard penetration test (SPT), cone
penetration test (CPT) and shear wave velocity measurements.
Robertson and Wride (1998) proposed a method based on CPT data. In
their method, CRR
M=7.5
(related to a reference earthquake magnitude of
M =7.5) can be evaluated from the following simplied equations:
CRR
M=7.5
= 93
_
(q
c1N
)
cs
1000
_
3
+ 0.08 (4.60)
if 50 (q
c1N
)
cs
160.
CRR
M=7.5
= 0.833
_
(q
c1N
)
cs
1000
_
+ 0.05 (4.61)
if (q
c1N
)
cs
< 50.
In the equations, (q
c1N
)
cs
is the clean sand equivalent of the stress-
corrected cone tip resistance which is a function of q
c
and sleeve friction
f
s
. In the Robertson and Wride model, a comprehensive procedure is used to
determine (q
c1N
)
cs
through the use of some intermediate parameters, includ-
ing the soil behavior index I
c
as dened already by Robertson for identication
of soils out of CPT.
Finally, a factor of safety against liquefaction for an earthquake magnitude
M can be evaluated as:
FoS =
CRR
M=7.5
CSR
MSF
where MSF is the magnitude scaling factor to convert the CRR
M=7.5
to the
equivalent CRR for the design earthquake. The recommended MSF for this
CPT based method is MSF = 174/M
2.56
.
Slope stability analysis of the Doeldok embankment 51
4.7 Slope stability analysis of the Doeldok embankment
In the analysis of stability of the Doeldok embankment, the Strength Re-
duction method was mainly used by means of a nite element based program
(PLAXIS). Nevertheless, the method of slices was used here as well to conrm
the outcome. Soil properties for each soil type in the analysis can be found in
section 6.3.4.
In the stability analysis no account was taken of the special procedure for
sensitive clays since none of the soils involved classied as highly sensitive.
However, the risk of liquefaction was assessed using the method proposed by
Robertson and Wride (1998).
For characterizing the local seismicity at the Doeldok area, an earthquake
magnitude of M =5.5 was assumed and a Peak Ground Acceleration (PGA)
of 0.05g was obtained from the seismic zonation map of Belgium, see gure
4.19. The embankment is located in the harbor of Antwerp, nearby point An
in the gure that falls in Zone 1.
Making use of the latest data of CPT tests performed under water through
the sand body of the embankment at dierent locations, a factor of safety
could be evaluated (see Figs 4.20, 4.21, 4.22, 4.23 and 4.24). In all cases the
safety factor against liquefaction did exceed 1, in fact most factors ranged from
FoS = 2.5 to 6. One can conclude that liquefaction within the embankment
body for an earthquake magnitude of 5.5 will not occur.
The outcome of the analysis of stability of the Doeldok embankment is
given in more detail in chapter 7.
Figure 4.19. Seismic zonation of Belgium (NBN-ENV 1998-1-1: 2002 NAD-
E/N/F)
52 Discussion on slope stability evaluation
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
1
5
q
c

(
M
P
a
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
2
3
4
5
F
r
i
c
t
i
o
n

r
a
t
i
o

F

(
%
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
0
1
0
0
N
o
r
m
a
l
i
s
e
d

q
c

(
q
c
1
N
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
0
.
2
0
.
4
C
S
R
,

C
R
R
TAW level
C
S
R
C
R
R
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
F
o
S

a
g
a
i
n
s
t

l
i
q
u
e
f
a
c
t
i
o
n
TAW level
Figure 4.20. Liquefaction assessment out of CPT of the embankment body above
the SSI improved zone (CPT3)
Slope stability analysis of the Doeldok embankment 53
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
1
5
q
c

(
M
P
a
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
2
3
4
5
F
r
i
c
t
i
o
n

r
a
t
i
o

F

(
%
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
0
1
0
0
1
5
0
N
o
r
m
a
l
i
s
e
d

q
c

(
q
c
1
N
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
0
.
2
0
.
4
C
S
R
,

C
R
R
TAW level
C
S
R
C
R
R
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
F
o
S

a
g
a
i
n
s
t

l
i
q
u
e
f
a
c
t
i
o
n
TAW level
Figure 4.21. Liquefaction assessment out of CPT of the embankment body above
the SSI improved zone (CPT5)
54 Discussion on slope stability evaluation
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
1
5
q
c

(
M
P
a
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
2
3
4
5
F
r
i
c
t
i
o
n

r
a
t
i
o

F

(
%
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
0
1
0
0
1
5
0
N
o
r
m
a
l
i
s
e
d

q
c

(
q
c
1
N
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
0
.
2
0
.
4
C
S
R
,

C
R
R
TAW level
C
S
R
C
R
R
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
F
o
S

a
g
a
i
n
s
t

l
i
q
u
e
f
a
c
t
i
o
n
TAW level
Figure 4.22. Liquefaction assessment out of CPT of the embankment body above
the non-improved zone (CPT8)
Slope stability analysis of the Doeldok embankment 55
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
1
5
q
c

(
M
P
a
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
22
3
4
5
F
r
i
c
t
i
o
n

r
a
t
i
o

F

(
%
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
0
1
0
0
N
o
r
m
a
l
i
s
e
d

q
c

(
q
c
1
N
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
0
.
2
0
.
4
C
S
R
,

C
R
R
TAW level
C
S
R
C
R
R
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
F
o
S

a
g
a
i
n
s
t

l
i
q
u
e
f
a
c
t
i
o
n
TAW level
Figure 4.23. Liquefaction assessment out of CPT of the embankment body above
the non-improved zone (CPT14)
56 Discussion on slope stability evaluation
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
1
5
q
c

(
M
P
a
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
2
3
4
5
F
r
i
c
t
i
o
n

r
a
t
i
o

F

(
%
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
1
0
0
2
0
0
N
o
r
m
a
l
i
s
e
d

q
c

(
q
c
1
N
)
TAW level
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
0
.
2
0
.
4
C
S
R
,

C
R
R
TAW level
C
S
R
C
R
R
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
5
1
0
F
o
S

a
g
a
i
n
s
t

l
i
q
u
e
f
a
c
t
i
o
n
TAW level
Figure 4.24. Liquefaction assessment out of CPT of the embankment body above
the non-improved zone (CPT21)
5
Evaluation of consolidation
W.F. Van Impe, P.O. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
5.1 One-dimensional consolidation theory
Saturated deposits of low permeability soils, when loaded by a surcharge
(i.e. construction of an embankment on saturated clay), undergo settlements
occurring over a long period of time. This phenomenon is called consolida-
tion and it was extensively studied by Terzaghi in 1914 by means of a simple
piston-spring-water model (Fig. 5.1).

Clay

Spring
Water
Figure 5.1. Terzaghis piston-spring-water model
Afterward, the consolidation theory was extended and improved in many
ways (e.g., Barden, 1965; Davies et al., 1965; Gibson et al., 1967; Schiman,
1980; Gibson et al., 1981). Nowadays, two theories can be well dierentiated:
infinitesimal strain and finite strain theories of consolidation.
The innitesimal strain theory assumes that the deformation of the soil
mass is so small that the compressibility and the hydraulic conductivity
remain constant for a given load increment. This assumption was shown to
produce good results for many practical problems. However, when very soft
soils are involved wherein the properties (compressibility and permeability)
gradually change with consolidation progress, such assumption may lead to
erroneous predictions.
Recognition of the limitations of small strain consolidation theories led to
the development of large or nite strain models in which no restrictions are
imposed on the deformation of the compressible media. However, the price
58 Evaluation of consolidation
one has to pay for a better approach is a more complex procedure that can
be solved only numerically.
General and common assumptions usually accepted in the derivation are:
The porous medium consists of incompressible pore water, incompressible
mineral particles and deformable skeleton.
The deformation consists of the rearrangement of mineral particles accom-
panied by a ow of pore water. The volume of soil solids remain constant.
Skeleton is homogeneous; i.e., a single void ratio-eective stress and void
ratio-hydraulic conductivity relationships govern the entire soil mass.
The clay layer is normally consolidated.
The water ow is one dimensional, and it is motivated by mechanical
forces such as surcharge loading. Thermal, electrical or chemical potential
to induce ow are not included.
Eective stress principle applies.
The ow of uid through the porous skeleton is governed by the linear
Darcy-Gersevanov relationship.
The next section describe each theory and solutions in more detail.
5.2 Innitesimal strain theory
As initiated by the early work of Terzaghi on stress strain analysis of saturated
soils, the one-dimensional small strain consolidation theory can be expressed
in a rather easy way.
Lets consider a saturated element of porous material with characteristics
fullling the assumptions above. In order to obtain the equation governing
the consolidation, we start from the continuity requirement for the solid and
liquid interaction. The continuity equation for the solid phase requires:
[(1 n)
s
v
s
]
z
+
[(1 n)
s
]
t
= 0 (5.1)
In the same way, the continuity equation for the uid requires:
[n
w
v
w
]
z
+
[n
w
]
t
= 0 (5.2)
where v
s
and v
w
are the absolute velocities of the solid and liquid phase
respectively in an Eulerian coordinate system;
s
and
w
are the unit weight
of the solid and uid phase; n is the porosity of the element, z is the vertical
coordinate and t is the time.
Combining equations 5.1 and 5.2, the continuity condition for the mixture
will be:
[(1 n)v
s
+nv
w
]
z
= 0 (5.3)
Innitesimal strain theory 59
Considering that Darcys law still governs the ow of water through the
soil skeleton (i.e. accepting all boundary conditions limited to a Darcy type
of ow), the equilibrium equation of water can be expressed as:
u
z
+n

w
k
(v
w
v
s
) = 0 (5.4)
where u is the excess pore pressure and k is the hydraulic conductivity.
Dierentiation of 5.4 with respect to z leads to:

2
u
z
2


w
k
[n(v
w
v
s
)]
z


w
k
2
k
z
n(v
w
v
s
) = 0 (5.5)
and combining with equation 5.3, it gives:

2
u
z
2


w
k
v
s
z
+
1
k
k
z
u
z
= 0 (5.6)
The partial derivative v
s
/z can be expressed in terms of settlement
rate, in order to introduce the constitutive relationship for the soil skeleton,
therefore, the following holds true:
v
s
z
=

1
t
= m
v

v
t
(5.7)
and the previous equation (5.6) can be written as:

2
u
z
2
+

w
k
m
v

v
t
+
1
k
k
z
u
z
= 0 (5.8)
The change of vertical eective stress with time can be expressed in terms
of change of pressure applied at the surface and change of excess pore water
pressure:

v
t
=
q
t

u
t
(5.9)
and by substituting 5.9 in 5.8, the general equation of one-dimensional con-
solidation can be expressed in terms of excess pore pressure:

2
u
z
2
+m
v

w
k
_
q
t

u
t
_
+
1
k
k
z
u
z
= 0 (5.10)
Now, If we assume that the hydraulic conductivity remains constant, then
equation 5.10 reduces to:
c
v

2
u
z
2
=
u
t

q
t
(5.11)
where c
v
= k/(m
v

w
), is the coecient of consolidation and the term q/t
describes the change of surcharge load with time (variable load).
Then, equation 5.11 allows modeling of consolidation due to variable load-
ing (i.e. staged construction) when changes of hydraulic properties during
loading are neglectable.
60 Evaluation of consolidation
h
0
h
1
h
L
h
n
h
n+1
H
q(t)
IMPEDANCE LAYER k
0
IMPEDANCE LAYER
z
zr
zr+1
k
n+1
cv
1
, k
1
, mv
1
cv
L
, k
L
, mv
L
cv
n
, k
n
, mv
n
STARTING VALUES (T=0) BOUNDARY CONDITIONS
BOUNDARY CONDITIONS

n
TIME
DEPTH
FREE DRAINING BOUNDARY
FREE DRAINING BOUNDARY
Figure 5.2. Multi-layered system and nite dierence solution scheme
5.2.1 Numerical solution
The one-dimensional consolidation is governed by the parabolic partial dier-
ential equation 5.11. The problem is classied as one of initial value since we
know in advance the excess of pore water pressure at a time t
0
= 0 and we
want to estimate a new value at a time t = t
0
+ t. The problem can also
be extended to allow several layers of clay with dierent properties, a multi
layered system (Fig. 5.2).
It is assumed that the soil prole consists of n contiguous layers. The
arbitrary layer is indexed L with thickness h
L
. The soil properties of the L
th
layer are the coecient of consolidation c
L
v
, the compressibility m
L
v
, and the
coecient of permeability k
L
. The compressible stratum is the system of n
compressible layers and has a total thickness H.
The analysis that follows describes the solution proposed by Schiman and
Arya (1977).
Indexes which refer to a layer are written as superscripts. The superscript
L refers to an arbitrary layer. Thus, all soil properties are superscripted. The
space coordinate z is a global coordinate and has its origin at the surface
z = 0 (Fig. 5.2). All indexes that depend on the global space coordinate z
are written as subscripts. Similarly, all indexes that are dependent of time are
written as subscripts.
As a general rule the superscript L refers to a layer number; the subscript
i refers to a z coordinate point and the subscript j refers to a value at a
particular time. The value of the superscript L runs from 1 to n. The value of
i runs from 0 at the surface (z = 0) to a value p at the lower boundary (z = H).
Innitesimal strain theory 61
The subscripted space index at the layer interface is r. The subscripted time
index j runs from 0 at t = 0 in an arithmetic progression (0, 1, 2, 3).
The consolidation phenomena of a single layer is governed by equation
5.11 alone. However, in a multi-layered system there are n similar equations
which must be solved in order to determine the excess pore water pressure at
any point in space and time. Equation 5.11 can be rewritten as:
c
L
v

2
u
L
z
2
=
u
L
t

dq
dt
L = 1, 2, ..., n (5.12)
The three types of time-independent boundary conditions that can apply
to the stratum boundaries z = 0 and z = H can be expressed in general form
as:
a
1
u
1
z
(0, t) b
1
u
1
(0, t) = c
1
(5.13)
a
n
u
n
z
(H, t) b
n
u
n
(H, t) = c
n
(5.14)
where the coecients a
1
, b
1
, c
1
, a
n
, b
n
and c
n
take on specic values for specic
conditions (i.e. free draining, impervious and partial drainage or impeded).
Table 5.1 presents the particular values of these coecients.
Table 5.1. Boundary conditions
Boundary Upper boundary Lower boundary
condition a
1
b
1
c
1
a
n
b
n
c
n
Free draining 0 1 0 0 1 0
Impervious 1 0 0 1 0 0
Impeded h
1

1
=
k
0
h
1
k
i
h
0
0 h
n

n
=
k
n+1
h
n
k
n
h
n+1
0
In addition to the boundary conditions for the compressible stratum, it
is assumed that there is full continuity between clay layers. This assumption
requires that the excess pore water pressure and ow velocities in adjacent
layers are equal at the common layer interfaces. It gives:
u
L
(z
r
, t) = u
L+1
(z
r
, t) and k
L
u
L
z
(z
r
, t) = k
L+1
u
L+1
z
(z
r
, t) (5.15)
where, as shown in gure 5.2, the distance z
r
is the distance from the surface
to the layer interface separating the L
st
and the (L+1)
st
layer. So far, all the
formulation has been described.
A solution can be applied at this point making use of either nite dierence
methods or nite elements. The nite dierence procedure was chosen here.
The nite dierence method makes use of a discretization of the space and
62 Evaluation of consolidation
time and therefore replaces the continuous derivatives by the ratio of changes
in the variable over a small but nite increments.
A program, SSCON-FD (Small Strain CONsolidation Finite Dierence
solution), has been developed here on the basis of the previous procedure. The
program was written in Maple. As explained before, it allows the innitesimal
strain consolidation analysis of a multilayered soil deposit. In the model each
layer may have dierent consolidation parameters and the boundary condi-
tions allow not only fully drained or fully closed conditions but also partial
drainage by introducing hydraulic conductivity values of impedance layers
(see Fig. 5.2).
5.2.2 Applications of SSCON-FD
In this section the range of applications of the program are described, with
special attention to the analysis of staged loading. For this purpose, 2 ctitious
problems have been addressed and numerical solutions obtained.
[Link] Problem 1
Lets assume that a sand embankment is constructed in stages over a satu-
rated clay layer. The clay layer has a thickness of 5 m, c
v
=3 m
2
/year and
k =0.02 m/year, it rests over a silty layer with a thickness of 1 meter with
k =0.03 m/year. The silty layer rests over a deep sand with high permeability.
The rst load stage (80 kPa) was placed uniformly over a period of 5 months.
The work was stopped for 1 year and nally, the second load stage was applied
with the same magnitude and at the same rate.
It is required to study the consolidation of the clay layer:
[Link] Solution
SSCON-FD was used to solve the problem. Free draining conditions were
adopted for the upper boundary and impeded conditions for the lower since
there might exist downwards ow. Results are shown in gures 5.3 and 5.4.
Figure 5.3 illustrates the development of eective stress with time (at the
midheight of the layer) and gure 5.4 describes the excess pore pressure in
time and space.
[Link] Problem 2
An instantaneous load of 100 kPa has been applied at the top of a 4-meter
thick soil deposit that rests over a sound rock with very low permeability.
The soil deposit consists of two dierent clays layers with the same thick-
ness. The consolidation properties of the upper clay are: c
v
=1.5 m
2
/year
and k =0.4 m/year. The consolidation properties of the lower clay layer are:
c
v
=0.2 m
2
/year and k =0.1 m/year.
It is required to study the consolidation of the layered system.
Innitesimal strain theory 63
3 0 j .02
.
0 0.5 1 1.5 2 2.5 3
0
50
100
150
200
Excess PWP at midpoint
Excess PWP at bottom
Total stress increment
Effective stress increment at midpoint
TIME, [year]
Figure 5.3. Consolidation analysis of staged loading (SSCON-FD)
Z
0 0.5 1 1.5 2 2.5 3
95
90
85
80
80
75
75
75
70
70
70
65
65
65
65
60
60
60
60
60
55
55 55
55
55
50
50
50
50
50
45
45
45
45
45
45
40
40
40
40
40
40
35
35
35
35
35
35 30 30
30
30
30
30 25
25 25
25
20 20
20
15 15
15
Figure 5.4. Contour lines of excess pore water pressure (SSCON-FD) as a function
of space (vert. axis) and time(horiz. axis)
[Link] Solution
Again, SSCON-FD was used to solve the problem. Free draining conditions
were adopted at the top boundary and impermeable at the bottom. The results
are shown in gure 5.5. From the excess pore water pressure proles, numer-
ically predicted up to 3 years, one can clearly note that the one-dimensional
consolidation of the upper layer goes faster than in case of the lower layer due
to the considerable dierence in consolidation parameters.
Since SSCON-FD allows the modeling of multilayered systems, thick layers
of clay showing dierent parameters with depth could be modeled as such,
64 Evaluation of consolidation
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 20 40 60 80 100
Excess pore water pressure, kPa
D
e
p
t
h
,

m
T=0.0
T=0.5
T=1.0
T=1.5
T=2.0
T=2.5
T=3.0
Figure 5.5. Excess pore water pressure proles (SSCON-FD)
in this way, avoiding having to rely on an average consolidation parameter
for the whole strata. Such practice usually leads to erroneous predictions as
demonstrated by Pyrah (1996).
5.3 Finite strain theory
General three-dimensional consolidation approaches have been discussed al-
ready in literature starting from the important work of Biot (1941); however,
all of it related to small strain levels of deformations of the soil skeleton.
The most general theory of a one-dimensional type of consolidation is that
proposed by Gibson et al. (1967). This analysis overcomes the limitations that
the classical, small strain, theory entails; but at the same time the problem be-
comes so complex that only numerical solutions can be obtained for practical
problems.
The process of nite strain one-dimensional consolidation of a saturated
porous medium is governed by:

z
_
g(e)
e
z
_
b(e)
e
z
+
e
t
= 0 (5.16)
where
g(e) =
k(e)

w
(1 +e)
d

de
b(e) =
_

w
1
_
d
de
k(e)
1 +e
in which e is the void ratio,
s
and
w
are the solid and uid weights per unit
of their own volume, respectively, and z is a reduced coordinate encompassing
a volume of solids in a volume of unit cross sectional area lying between the
datum plane and the Lagrangian coordinate point (Gibson, 1967).
Finite strain theory 65
The function g(e) plays the role of consolidation coecient and b(e) intro-
duces the eect of gravity. If the gravity eect is neglected [i.e. b(e) = 0] and
g(e) is assumed to remain constant during the process, then equation 5.16
would simplify to the classical theory (i.e. Terzaghis).
Equation 5.16, with appropriate boundary and initial conditions and con-
stitutive properties, provides the governing relationship from which a solution
can be developed. The required constitutive properties are: the relationship
between void ratio and eective stress, and the relationship between coecient
of permeability and void ratio.
It is noted that the governing equation, while unrestricted as to the mag-
nitude of strain and non linearity of constitutive relationships, it is based
upon the premise of homogeneity and monotonic behavior; load-unload-reload
cycles are not permissible.
Many researchers have attempted to solve the problem (Cargill, 1984; Fox
and Berles, 1997; Van Impe P.O., 1999). The work of Van Impe P.O. (1999)
extended the formulation introducing the sedimentation phase of a soil deposit
during its skeleton formation. Within this framework, two programs were
elaborated: FISCC (Van Impe P.O., 1999) and CBFISCC. Both analyze the
one dimensional consolidation of a single clay layer subjected to the action of
an instantaneous constant load introducing dierent numerical methods for
the solution of the governing dierential equation.
CBFISCC is a numerical solution of the partial non-linear dierential equa-
tion governing the phenomena. The equation is transformed to a system
of ordinary dierential equations and the system is solved using an appro-
priate algorithm.
FISCC is piecewise linear model. It encompasses an iterative procedure
in which each step is linearly evaluated, but it becomes non linear in
the overall solution. To apply this technique it is essential to use su-
ciently small time steps; therefore, it obviously requires more time for
computation.
The input data of such programs usually consist of:
initial thickness of the compressible layer
specic gravity of solids and water
initial void ratio of the compressible layer
boundary drainage conditions (open or close)
total stress increment at the surface
constitutive relationships of the soil (e.g. K=K(

v
) and e =e(

v
))
The program output consist of:
settlement versus time
void ratio proles at several time steps
pore water pressure proles at several time steps
66 Evaluation of consolidation
Degree of consolidation
(Finite strain theory)
Degree of settlement
(Finite strain theory)
Degree of consolidation
(Infinitesimal strain theory)
Time [years]
D
e
g
r
e
e

o
f

s
e
t
t
l
e
m
e
n
t

a
n
d

c
o
n
s
o
l
i
d
a
t
i
o
n
1000 100 10 1
1.0
0.8
0.6
0.4
0.2
0.0
Figure 5.6. Degrees of settlement and consolidation (pore pressure dissipation)
(Schiman, 1994)
5.4 Innitesimal strain versus nite strain theory
As demonstrated in the previous section the conventional one-dimensional
consolidation theory is nothing but a special case of the more general theory
formulated by Gibson et al. (1967).
Figure 5.6 shows the results of a comparative study performed for a 10 m
thick, normally consolidated St. Herblain clay layer subjected to a surcharge
of 200 kPa (Schiman, 1994).
In nite strain consolidation analysis, unlike in innitesimal theories, there
are two measures of the progress of consolidation: the degree of settlement
(ratio of settlement at a time t to the nal settlement) and the degree of pore
water pressure dissipation.
As shown in gure 5.6, the conventional (innitesimal) theory underpre-
dicts the rate of settlement and overpredicts the rate of pore water pressure
dissipation. Clearly, the use of nite strain theory would provide a better
(safer) estimates of the shear strength gain of a consolidating clay.
5.5 Consolidation at the Doeldok site
As far as the actual case of the embankment foundation soil is concerned, a
simulation was performed to evaluate both, the progress of settlements and
Consolidation at the Doeldok site 67
0
5
10
15
20
25
30
35
40
45
0 0.5 1 1.5 2
Time (year)
A
v
e
r
a
g
e

c
o
n
s
o
l
i
d
a
t
i
o
n

d
e
g
r
e
e

(
%
)
From Finite strain theories (settlements)
From Finite strain theories (PWP)
From Infinitesimal strain theories
Infinitesimal strain
solutions
Finite strain solution
(settlement)
Finite strain solution
(pore water pressure)
Figure 5.7. Consolidation analysis of the foundation soil evaluated with large and
small strain theories
pore water pressure dissipation. In this simulation, both theories of consoli-
dation (large strain and small strain) were applied implementing the consti-
tutive equations (K=K1(

v
) and K=K2(e)) for the soft dredged material
obtained from several tests (see section [Link]). In this simulation, a sin-
gle load increment equal to the weight of the whole embankment has been
studied. The results are illustrated in gure 5.7 which conrm that the excess
pore water pressure dissipation and the settlement will take place at dierent
rates. The output of small strain analysis in gure 5.7 is showed as a range
because there is a range of consolidation coecients that can be chosen out
of the constitutive equations.
Clearly, the pore water pressure is expected to dissipate quite slowly (after
2 years only 15% of the total excess will dissipate) while signicant settlements
are expected to be observed in the same time period (after 2 years already 40%
of the total settlement will occur). Moreover, it can be concluded that small
strain theories overestimate the degree of consolidation from excess pore water
pressure but they underestimate the degree of consolidation from settlements.
As it is discussed in chapter 7, the presence of the soft soil as foundation
soil for the underwater embankment required a controlled rate of construction
in the framework of a staged construction. Staged construction allows the
foundation soil to partially dissipate its excess pore water pressure and relies
on a strength increase that depends on the consolidation degree. This means
that measuring or estimating the consolidation degree of the foundation soil
during construction is of utmost importance. Therefore, the foundation soil
was provided with a number of pore water pressure transducers at several
location and at 3 dierent levels within the layer. Moreover, settlement proles
68 Evaluation of consolidation
0
20
40
60
80
100
120
140
160
180
200
0 100 200 300 400 500 600 700
Time (days)
E
x
c
e
s
s

P
o
r
e

w
a
t
e
r

p
r
e
s
s
u
r
e

(
k
P
a
)
Construction
works
(Phase 1)
Consolidation
Excess pore water pressure
in the soft soil deposit
under the embankment
Figure 5.8. Excess pore water pressure at one location in the foundation soil
under the embankment were also regularly measured by means of a clever
and simple system that makes use of a plastic tube that is laid under the
embankment and that is lled with water. Then, settlement can be derived
out of water pressure (head) measurements inside the tube by means of a
probe that is pulled along the plastic tube. Those 2 instrumentation means
allowed for a controlled construction.
Figures 5.8 and 5.9 illustrate the progress of excess pore water pressure
and settlements under the embankment load during construction up to the
current situation. A more detailed explanation of the construction proce-
dure can be found in chapter 9, but in a nutshell, the rst phase of the
dam was built in stages allowing for some time in between. At the end
of Phase 1 an even longer period was introduced to allow the soft soil to
consolidate.
The results of the monitoring of pore water pressures and settlements do
show indeed that there has been a very slow dissipation of pore water pressure
and a much faster progress of settlements.
A more elaborated analysis of the consolidation progress at the Doeldok
site under the current loading situation was attempted here and results were
compared to measurements (see Figs 5.10 and 5.11). As expected from the
previous discussions it can be observed that the large strain consolidation
prediction match the measurements more accurately than small strain predic-
tions. The small deviation observed may just be consequence of the natural
inhomogeneity of the soft foundation soil.
Consolidation at the Doeldok site 69
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0 100 200 300 400 500 600 700
Time (days)
S
e
t
t
l
e
m
e
n
t

(
m
)
Settlement of the soft
soil deposit under the
embankment
Figure 5.9. Settlements of the foundation soil under the embankment
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700
Time (days)
D
e
g
r
e
e

o
f

c
o
n
s
o
l
i
d
a
t
i
o
n

f
r
o
m

s
e
t
t
l
e
m
e
n
t
s

From measurements
From small strain theories
From large strain theories
Figure 5.10. Degrees of consolidation of the foundation soil studied here out of
settlements
70 Evaluation of consolidation
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700
Time (days)
D
e
g
r
e
e

o
f

c
o
n
s
o
l
i
d
a
t
i
o
n

f
r
o
m

e
x
c
e
s
s

P
W
P
From measurements
From small strain theories
From large strain theories
Consolidation
Construction
works
(Phase 1)
Figure 5.11. Degrees of consolidation of the foundation soil studied here out of
pore water pressure
5.6 Conclusions
The nite strain theory constitutes outspokenly a better approach to the
phenomena of consolidation. It was shown that the use of innitesimal
strain consolidation theories may be underconservative for staged construction
analysis.
Nevertheless, at the very rst stage of design, the classical theory may
still provide the basis for further renement of calculations. Clearly, many
geotechnical engineers are more familiarized with consolidation parameters
for the classical (innitesimal strain) theory than the nite strain theory. As
a rule, the softer the soil the larger the error introduced in the analysis by
not taking into account changes of compressibility and hydraulic conductivity
during consolidation.
6
Geotechnical characterization of the site
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
J. Van Mieghem & A. Baertsoen
Ministry of Flanders, Belgium
6.1 Overview
The present chapter illustrates the geotechnical characterization of the soils
present in the site. Also some estimations are proposed of the expected prop-
erties of the sand embankment body.
The site of investigation is a dock located in the harbor of Antwerp (Bel-
gium). The water depth to the sediments level is about 20 m in the area.
6.2 Soil prole and characterization
In the framework of a geotechnical investigation of the foundation soil at the
site, a number of eld tests such as CPT and eld vane tests were carried out
focusing on the characterization of the soft material. Figure 6.1 illustrates the
location of the tests in the area. Moreover, several borings were performed
to collect soil for laboratory testing. All eld tests were carried out from a
jack-up platform.
Figure 6.2 illustrates a typical CPT prole. Clearly the soft material (with
q
c
<0.3 MPa) extends for about 8 m. Underneath, a relatively thin (thickness
of 2 m) sand layer can be found resting on a deep highly overconsolidated clay
from the Tertiary (Boom clay).
Out of the eld vane testing it was possible to estimate not only the
undrained shear strength at several depths but also to conrm the normal
consolidation state of this young deposit. Figure 6.3 illustrates the measured
undrained shear strength c
u
plotted against the estimated vertical eective
stress. With the exception of some points, the general trend seems to be linear
increasing with depth (with c
u
0.3

v
) as it would be expected in young
normally consolidated soils. Points deviating from the trend in the upper part
of the layer suggest the presence of a lightly overconsolidated crust; on the
other hand, points deviating from the trend in the lower layer may actually
belong to a transition zone of soft material and sand.
72 Geotechnical characterization of the site
C
A
A
B
B
C
D
D
0 50 100 150 250m 200
BEVEREN
DOELDOK
S7+V7
S2+V2
S3
S4
S5
S6+V6
S8
S9
S10
S11+V11
S1
S12
S20
VC17+V17
VC14+V14
S13
S15
S18
VC16+V16
VC19+V19
V1,V2,Field Vane tests
S1, S2, Cone penetration tests
VC1, VC2, Borings
Figure 6.1. Location of eld tests
-28
-26
-24
-22
-20
-18
-16
0 5 10
qc (MPa)
T
A
W

l
e
v
e
l
Sediments level
CPT on the natural
foundation soil
Soft dredged material
Sand
Boom clay
Highly overconsolidated
Figure 6.2. Typical CPT prole in the area
6.3 Selection of parameters for design
The design of the embankment involves dierent materials such as: sand (for
the embankment itself), Boom clay and dredged sludge (with and without
improvement). Each material and its properties are described below.
Selection of parameters for design 73
0
2
4
6
8
10
12
14
0 2 4 6 8 10
c
u
(kPa)

'
v

(
k
P
a
)

c
u
= 0.3 '
v
From field vane testing
OC upper crust
NC state
NC layer
Soil in
OC state
Soil in
Figure 6.3. Field vane test results
6.3.1 Dredged material
The soft dredged material is actually the result of years of sedimentation
and self-weight consolidation of dredged sediments from waterways within
the harbor of Antwerp. Although there were attempts in the past to improve
this material by vacuum consolidation, its consistency remained soft.
[Link] Physical properties
Table 6.1 summarizes some averaged physical properties of the material in its
natural state. The natural water content of the soil is of the order of 115%,
the plasticity index of the order 77 and the organic content of about 6%. pH
measurements of the pore water give a value close to 7.
In order to get a closer look of this material, a specimen of untreated
dredged material was analyzed on the Scanning Electron Microscope (SEM).
The scanning electron microscope is a type of electron microscope capable
of producing high resolution images of a sample surface at magnication levels
that could go up to molecular levels. These images have a characteristic 3D
appearance and are useful for judging the microstructure of a sample.
The working principle of SEM is simple, electrons are shot to a sample
and as a result of their interaction (electrons and matter) electrons from the
sample are released. Such released electrons provide information of the sample,
74 Geotechnical characterization of the site
Table 6.1. Physical properties of the dredged material
Index Value
Liquid limit 124.4
Plastic limit 46.7
Plastic index 77.7
Natural water content, % 115.0
Organic content, % 6.0
Natural lime content, % 13.9
Sand fraction, % 10.4
Wet density, g/cm
3
1.4
pH of the pore water 7.2
Electron beam
BSE
SE
P
X
-
r
a
y

d
e
t
e
c
t
o
r
S
E

d
e
t
e
c
t
o
r
X-ray diffraction SE (or BSE) imaging
Vacuum chamber
SE: Secondary electrons
BSE: Back-scattered electrons
P: Photons
S
a
m
p
le
Figure 6.4. Scanning electron microscopy working principle
therefore they are detected and analyzed (Fig. 6.4). The primary electron
beam is produced by an electron gun and accelerated toward the sample.
The sample must be conductive so that the energy of the inciding electrons
can be diverted from the sample, otherwise, the electron beam could burn and
destroy the sample before any analysis can be done. That is not an issue when
analyzing steel samples, but in the case of soils (non conductive) the samples
should be coated with a ne (nano scale) layer of a conductive element such
as gold. Moreover, samples are tested dry in a vacuum chamber to minimize
any interaction of the inciding and emitted electrons (from the sample) with
air particles. However, newly developed SEM allow testing samples with some
amount of water and coating is no longer mandatory.
Selection of parameters for design 75
Figure 6.5. Scanning electron microscopy analysis of natural dredged material
The electron beam is deected by scanning coils so that it can raster an
area of the sample surface. As the primary electrons strike the surface they
are elastically and inelastically scattered by atoms in the sample. Through
these scattering events, the primary electron beam eectively spreads and
lls a teardrop-shaped volume, known as the interaction volume, extending
about less than 100 nm to 5 m depths into the surface. Interactions in this
region lead to the subsequent emission of electrons and photons (e.g. X-rays)
which are then detected to produce an image. Two main types of emitted elec-
trons can be identied: backscattered electron (BSE) which are high energy
electrons product of elastic interaction of the electron beam with the sample
and secondary electrons (SE) which are low energy electrons from inelastic
interactions.
X-rays, usually detected in a SEM, allow the study of the chemistry
of a particular point in the surface of the sample by identication tech-
niques based on energy dispersive X-ray spectroscopy or wavelength dispersive
X-ray spectroscopy.
SE electrons (more commonly used for SEM imaging) are detected by a
scintillator-photomultiplier device and the resulting signal is rendered into
a two-dimensional intensity distribution that can be viewed and saved as a
digital image for each point on the sample surface that the electron beam
rasters. Finally, a sharp image of the surface can be generated.
The specimen analyzed under the SEM was prepared to simulate the natu-
ral in situ conditions of the dredged material with very light hand compaction.
Figure 6.5 does illustrate a SEM picture of the specimen at an amplifying fac-
tor of 1100. Clay particles (platy shaped) can be identied by their somewhat
more bright color in the picture. They seem to be uniformly spread and are
interacting with the silt and sand particles (in edgy or rounded shape and
76 Geotechnical characterization of the site
0
20
40
60
80
100
120
140
0 50 100 150 200 250 300
'
c,
kPa
c
u
,

k
P
a
TX-CIUC
TX-UIUC
c
u
= 0.49 '
c
Isotropic
consolidation
Figure 6.6. Undrained strength ratio from laboratory testing
much darker in color). It is also possible to nd organic matter and bio-
logical remains showing a very regular micro morphology, typical in marine
soils.
[Link] Mechanical properties
As illustrated in gure 6.3, the undrained strength of the dredged mate-
rial is quite low. It was found that c
u
ranges from 2 to 4 kPa. For the
design phase, an initial value of c
u
=3 kPa has been assumed as the most
representative.
Furthermore, laboratory tests have been focused on the evaluation of a
undrained strength ratio (S
0
= c
u
/

v
). From a series of triaxial CU tests
on isotropically consolidated specimens (Fig. 6.6), a value of S
0
0.49 was
found. This conrmed also the fact that this soil has a normalized behavior.
However, it is known from literature (Kousoftas, 1981) that isotropic con-
solidation may overestimate the actual S
0
. In fact, the undrained strength
ratio estimated out of eld vane tests is only of the order of 0.3. Then a
S
0
= 0.3 was chosen for the design.
As for the stiness of the material an estimation of the drained Youngs
modulus was attempted by means of the graph in gure 6.7.
Figure 6.7 illustrates a chart to correlate the elastic Youngs modulus of
natural clay from the undrained strength, the overconsolidation ratio and
the plasticity index. We should look at the OCR=1 level corresponding to
PI > 50, therefore, we obtain E

y
/c
u
150. This conservative ratio has been
assumed for natural dredged material and also as a reference for the dredged
improved material.
In conclusion, the parameters adopted for natural dredged sludge are: c
u

3 kPa, c

= 0,

= 18

, = 12.8 kN/m
3
, and a drained Youngs modulus of
E

50
= 0.5 MPa.
Selection of parameters for design 77
Figure 6.7. Empirical correlation between E

y
/c
u
, OCR and PI (USACE, 1990)
[Link] Consolidation properties
The consolidation behavior of the dredged material has been assessed by
means of Constant Rate of Strain (CRS) tests, hydraulic conductivity tests
and oedometer tests. Figures 5 and 6 summarize the results of all tests per-
formed. A best tting curve have been drawn on each graph; in this way
the following constitutive equations necessary for large strain consolidation
analysis have been evaluated:
k = 6.0 10
8

1.18
v
(6.1)
k = 6.0 10
12
e
5.52
(6.2)
where the eective stress (

v
) and the permeability (k) are expressed in kPa
and m/s respectively.
[Link] Parameters of deep mixing improved zones
As far as strength of DM improved sludge is concerned, one may expect to
obtain a cone resistance varying from q
c
= 2 to 3.5 MPa (Van Impe W.F.,
2000) for a cement content of 5% to 15%. Therefore, with c
u
q
c
/15 (Nuyens
et al., 1995), a conservative mean undrained strength for two levels of improve-
ment (full and partial) would be about 100 kPa and about 50 kPa respectively.
These are not prediction but design values (minimum values) that will have
to be realized in the eld.
Then, parameters adopted for partially improved sludge are as follows:
c
u
50 kPa, c

= 0,

= 25

, = 13.5 kN/m
3
, and E

50
= 7.5 MPa.
78 Geotechnical characterization of the site
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
0 20 40 60 80 100 120 140
consolidation stress (kPa)
p
e
r
m
e
a
b
i
l
i
t
y

(
m
/
s
)
K = 6x10
-8
'v
-1.1773
R
2
= 0.7572
Figure 6.8. Hydraulic conductivity vs. vertical eective stress
1.E-11
1.E-10
1.E-09
1.E-08
1.E-07
1.E-06
1.E-05
1.E-04
0 2 4 6 8 10 12 14
void ratio
p
e
r
m
e
a
b
i
l
i
t
y

(
m
/
s
)
CRS
Oedometer
permeability test
K = 6x10
-12
e
5.5174
R
2
= 0.8812
Figure 6.9. Hydraulic conductivity vs. void ratio
6.3.2 Boom clay
The bottom layer on the soil prole consists of Boom clay. This material
has also been extensively studied in a number of papers (i.e. De Beer, 1967;
Bouazza et al., 1996).
De Beer (1967) studied the shear characteristics of the Boom clay for a
project of a tunnel under the Scheldt river in Antwerp, nearby the current
site of investigation.
According to a geological study carried out at that site, the Boom clay was
covered by about 40 m of Antwerpian sand at the beginning of the continental
Pleistocene erosion. This load acted on the Boom clay for 5 to 7 million years
and the unloading due to erosion started 500000 years ago. For that reason
Boom clay exhibits a brittle stress strain behavior, typical of overconsolidated
clay.
For the determination of undrained shear strength a large number of
unconned compression tests and undrained triaxial tests were performed
Selection of parameters for design 79
(De Beer, 1967) on samples extracted to depths of about 50 m. The following
relationship was deduced:
c
u
(kPa) = 75 + 3.5D(m) (6.3)
where c
u
is the undrained shear strength expressed in (kPa) and D is the
depth expressed in (m).
Simultaneously, in order to determine the drained shear strength parame-
ters a large number of consolidated undrained triaxial tests were carried out.
Due to considerable scatter, a range of friction angles was obtained, varying
from 17

to 24

and from 15

to 19

estimated with the peak and residual

strength respectively.
Bouazza et al. (1996) studied mechanical properties of reconstituted
Boom clay. The authors published stress-strain curves of shear tests from
which the undrained Youngs modulus has been estimated as round
E
u
=50 MPa and the drained Youngs modulus was estimated as
E

= 40 MPa.
Small strain stiness measurements (Haegeman, 1999) showed that the
Boom clay, under low levels of mean stress, has a E

max
(small strain Youngs
modulus) in the order of 80 MPa. Since the strain level for the problem
in consideration is certainly higher, the choice of drained Youngs modulus
E

50
=40 MPa seems to be reasonable.
Moreover, Bouazza et al. (1996) provide information about the intrin-
sic properties (critical state concept) of the Boom clay. They found that

cs
=18.5

.
Then, the following parameters have been adopted for Boom clay:
c
u
100 kPa, c

=0,

=19

, =19 kN/m
3
, and a drained Youngs modulus
of E

50
= 40 MPa.
6.3.3 Sand
As far as the embankment sand parameters are concerned, there are already
some experiences with sand dumping in harbor areas. For example, the con-
struction of the breakwater at the new outer harbor of Zeebrugge in Belgium.
A number of publications (De Wolf et al., 1983; Van Impe W.F., 1985; Van
Impe W.F., 1989) point up properties of this material.
The sand employed in that project consisted of rather coarse quartz parti-
cles with shells. The sand was used as replacing material in order to improve
the foundation of the breakwater. The sand was dumped with hopper suction
dredgers.
Here, it was assumed that a similar construction procedure would go on,
then, we can expect similar parameters. According to results obtained follow-
ing a quality control, the cone resistance in the sand after dumping varied
from 6 to 10 MPa corresponding to rather dense sand. Shear angles varying
in a range from 28

to 35

were reported.
80 Geotechnical characterization of the site
Figure 6.10. Young modulus vs. cone resistance for normally consolidated sands
(Van Impe W.F., 1986)
Moreover, we can also estimate deformation parameters from well estab-
lished correlations. Figure 6.10 allows correlating the Youngs modulus from
the cone pressure on NC sands.
In this way, the following parameters have been adopted here: c

= 0;

= 32

, =20 kN/m
3
,
d
= 17 kN/m
3
, and the drained Youngs modulus
E

y
15 MPa.
6.3.4 Summary of soil properties
Table 6.2 summarizes parameters adopted for design for each soil type.
Table 6.2. Soil parameters
Soil type c

c
u

d
E

50

(kPa) (

) (kPa) (kN/m
3
) (kN/m
3
) (MPa)
Natural dredged material 0 18 5 14 0.75 0.33
Improved dredged material 0 18 50 16 7.5 0.33
Sand 0 32 20 17 15 0.33
Boom clay 0 19 100 19 18 40 0.35
7
Design of underwater embankment of soft soil
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
H. Cecat & C. Boone
TECHNUM, Belgium
7.1 Overview
The present chapter illustrates the design of the underwater sand embankment
on soft soil (dredged material). The water depth to the sediments level is about
20 m. As detailed in the previous chapter, the soil prole at the site consists
of an upper soft layer overlying sand and a deep tertiary clay layer.
Several alternatives have been evaluated. More details one these alterna-
tives can be found elsewhere (Ver astegui, 2001). Figure 7.1 illustrates a scheme
of each of the options already studied:
Alternative A: The embankment implements full improvement at the toe
of the slope on the open river side while no improvement is specied for
the toe at the dry dock side.
Alternative B: The embankment comprises full improvement at the open
river side and partial improvement at the dry dock side.
Alternative C: In this conguration, the geometry slightly changes, partial
improvement of the soft material is considered on both toes.
Alternative D: The embankment has the same conguration as Alternative
C except for the absence of improvement at the toe on the dry dock side.
Alternative E: All the soft material is removed and the sand embankment
is directly founded on the Boom clay (Fig. 7.1). Technically this could be
qualied as a good solution, however, one must take into account that the
removal and new disposal of huge volumes of dredged material may be a
serious environmental issue.
The characteristics of the problem and the conditions of the foundation
soil demanded improvement of the foundation (by deep mixing techniques)
soil on the one hand and construction in stages on the other hand in order to
prevent early instability. Moreover, restrictions imposed on the total time of
construction led to adopt extra reinforcement elements such as geotextiles in
the embankment body.
82 Design of underwater embankment of soft soil
Boom clay
Natural sand Dredged sludge
F.I.
Sand embankment
Sand Backfill
Open river side Dry dock side
a) Alternative A
Sand embankment
Natural sand
F.I.
Sand Backfill
Dredged sludge
Boom clay
Partial Improvement
b) Alternative B
Partial Improvement Partial Improvement
Dredged sludge
Boom clay
Sand Backfill
Sand embankment
c) Alternative C
Partial Improvement
Sand embankment
Boom clay
Dredged sludge
Sand Backfill
d) Alternative D
Sand embankment
Sand Backfill
Dredged sludge
Boom clay
e) Alternative E
Figure 7.1. Previously studied design options
On the basis of the previous design alternatives, a more rened design has
been proposed (Fig. 7.2). A partial level of improvement is considered here at
the toes of the embankment.
Results of the design have shown that the construction can be safely carried
out in two main phases with a extended waiting time delay in between.
7.2 Geometry of the embankment
The general prole of the embankment and soil conditions are illustrated in
gure 7.2. As already described, the soil prole consists of about 8 m of soft
Stability analysis 83
PHASE 1
PHASE 2
4
10 16
4
Existing
sand dike
Boom clay
Dredged material IMP. Sludge [Link]
+11.0 TAW
-16.3 TAW
-24.3 TAW
-14.5 TAW
+4.0 TAW Dock side Open river side
65m 57m
Thin natural
sand layer
Figure 7.2. Embankment geometry nal design option
material, 2 m of sand and a deep highly overconsolidated clay (Boom clay).
The water depth to the sediments level is about 20 m.
The soft material was once subjected to vacuum consolidation to accelerate
the self-weight consolidation. That means that there exist drains installed
in the soft material that could help at accelerating the pore water pressure
dissipation. However, no real evidence exits to support such positive eect
of drains since there is no evidence that the drains are working well or are
clogged. Therefore all design that follows is based on the assumption that
there are no drains to be on the safe side.
Geotextile reinforcement was planned to be horizontally installed and dis-
tributed every 2 m on the face slope to the open river side as a protective
measure against wave action, to reduce the waiting time between phases and
to provide extra safety.
7.3 Stability analysis
In this section a detailed description of the design for the new alternatives is
given. PLAXIS (strength reduction method) and to a lesser extent SLOPE/W
(limit equilibrium method) were utilized to assess the problem.
7.3.1 Undrained analysis
A factor of safety of 1.3 was adopted here as minimum requirement to guar-
antee the safety during construction.
[Link] Phase 1
As illustrated in gure 7.2, the rst phase goes up to level +4.0TAW, that
means up to an embankment height of about 20 m. This phase was divided
in 10 sublayers to be constructed in stages every 2 months approximately to
allow for some consolidation and strength increase of the soft soil.
The following relationship (see also section 6.3.1 for the choice of param-
eters) was adopted to roughly estimate the undrained strength increase as
consolidation proceeds:
c
u
= 0.3

v
with

v
= U (7.1)
84 Design of underwater embankment of soft soil
0
2
4
6
8
10
12
0.00 0.15 0.30 0.45 0.60 0.75 0.90 1.05 1.20 1.35 1.50
Time [years]
A
v
e
r
a
g
e

C
o
n
s
o
l
i
d
a
t
i
o
n

D
e
g
r
e
e
,

%
Infinitesimal strain solution
Finite strain solution
Figure 7.3. Consolidation degree vs. time
Table 7.1. Increase of undrained shear strength for phase 1
Sublayer Level Time U
v
c
u
= c
uo
+c
u
TAW (year) (%) (kPa) (kPa)
1 14.3 0 0 20 3.00
2 12.3 0.16 4.7 40 3.28
3 10.3 0.33 5.8 60 3.71
4 8.3 0.50 6.6 80 4.20
5 6.3 0.66 7.3 100 4.75
6 4.3 0.83 7.8 120 5.35
7 2.3 1.00 8.3 140 5.99
8 0.3 1.16 8.7 160 6.66
9 +1.7 1.33 9.0 180 7.36
10 +4.0 1.5 9.4 220 8.09
where c
u
is the undrained strength gain,

v
is the eective stress incre-
ment,
v
is the total stress increment at the top of the layer, and U is the
average consolidation degree at a time t.
As discussed in chapter 5, there are two procedures to predict the con-
solidation behavior, the classical theory (innitesimal strain) and the nite
strain theory (large strain). Both methods were attempted here based on test
results and constitutive relations for consolidation given in chapter 6. Figure
7.3 illustrates for example the consolidation degree of the soft layer out of
the two analysis based on excess pore water pressure dissipation. The results
show that the large strain theory, which is a better approach to the actual
behavior, gives a smaller mean consolidation degree than the classical theory.
Then the use of the classical theory is underconservative in this case.
Table 7.1 summarizes all calculations of consolidation degree and undrained
strength increase for each sublayer of phase one calculated with the large
Stability analysis 85
3 kPa
3
.
2
8
k
P
a
C
u

=

3
.
7
1

k
P
a
.

4
.
2
0

k
P
a
.

4
.
7
5

k
P
a
.

5
.
3
5

k
P
a

5
.
9
9

k
P
a
6
.
6
6

k
P
a

7
.
3
6
k
P
a
8
.
0
9
k
P
a
-14.3TAW
-12.3TAW2
-10.3TAW
-8.3TAW
-6.3TAW
-4.3TAW
-2.3TA
-0.3TAW
+1.7TAW
+4.0TAW
TIME [days] 0 60 120 180 240 300 360 420 480
C
u

=

C
u

=

C
u

=

C
u

=

C
u

=

C
u

=

C
u

=

C
u

=

540
LEVEL
Figure 7.4. Construction program and available c
u
-3
-2
-1
0
1
2
3
4
5
1 1.2 1.4 1.6 1.8
Factor of Safety
L
e
v
e
l

T
A
W
Unsafe Safe
Figure 7.5. Factors of safety during construction of phase 1
Figure 7.6. Undrained stability analysis of just terminated phase 1
strain consolidation theory. Moreover, gure 7.4 illustrates more clearly the
construction program and the available c
u
strength of the soft layer for each
stage.
Figure 7.5 illustrates the results of the undrained stability analysis in terms
of factors of safety of the embankment following the construction program
showed in gure 7.4. Clearly, phase 1 can be built without concern about
failure as a factor of safety of 1.45 is reached at the end of construction of
phase 1 (Fig. 7.6).
86 Design of underwater embankment of soft soil
0
2
50
75
100
125
150
175
200
0 2 4 6 8 10 12
Time after Stage 1 [year]
T
d

[
k
P
a
]
Required minimum geotextile design
strength (Td) as a function of
consolidation time to safely
achieve phase 2 (with FS=1.3)
Figure 7.7. Geotextile design strength vs. waiting period between phase 1 and 2
Figure 7.8. Undrained stability analysis of reinforced phase2
[Link] Phase 2
As far as phase 2 is concerned, the undrained stability analysis showed that the
construction of phase 2 immediately after phase one would lead to instability.
Then a long waiting period should be allowed until the soft material increases
its strength up to a level enough to ensure safety.
Calculations showed that at least 13 years must be allowed. Given that
such long waiting period was not practical for this project, an alternative solu-
tion had to be proposed, that was geotextile reinforcement in the embankment
body.
As illustrated by gures 7.7 and 7.8, the geotextile reinforcement can in-
deed have a benecial eect to reduce the waiting period between phases 1
and 2 provided that at installation they are long enough to anchor themselves
in the sand and long enough to cross the potential slip surface.
For example, if a geotextile with a design strength of 50 kN/m is chosen,
then the waiting time required is about 8 years instead of 13 years.
7.3.2 Drained stability analysis
The embankment was then analyzed under drained conditions to evaluate
the stability in the long term. The results are illustrated in gure 7.9. As
expected, the long term stability of the embankment is not critical even when
Settlements 87
Figure 7.9. Drained stability analysis of nalized embankment without account of
geotextile reinforcement
6
0
VR(v)
VR(12)
VR(237)
1000 1 v
1 10 100 1 10
3
0
2
4
6
Vertical effective stress, kPa
V
o
i
d

r
a
t
i
o
Figure 7.10. Void ratio (VR) vs. vertical eective stress
the geotextile reinforcement is not considered. If we consider geotextile rein-
forcement, the factor of safety would increase from 1.34 to 1.42.
7.4 Settlements
Two approximations of settlements were made. The rst one makes use of
the compressibility parameters (i.e. compression index) estimated from the
constitutive properties (chapter 6) and the the second one, within the elasticity
framework, makes use of a reasonable Youngs modulus.
7.4.1 From Constitutive relationships
The maximum expected settlement was evaluated from the constitutive re-
lationships of consolidation behavior of the soft material: equations 6.1 and
6.2. Figure 7.10 illustrates the the relationship e

v
and the initial and nal
(after full construction) state of stress.
The maximum settlement is estimated with equation (7.2) where, H
0
is
the initial thickness of the layer, C is the compression index,

0
is the initial
vertical eective stress and is the total stress increment.
S

= H
0
1
C
ln
_

0
+

0
_
(7.2)
88 Design of underwater embankment of soft soil
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
Distance Section A-A [m]
S
e
t
t
l
e
m
e
n
t

[
m
]
Figure 7.11. Settlement prole at section A-A (from PLAXIS)
Then, we get:
S
max
3.0 m
7.4.2 From nite element program
The second method for settlement prediction is based on the theory of elastic-
ity. The term elastic settlement is usually associated to initial settlement
during undrained loading, however, elastic theory can also be used to estimate
settlements due to undrained loading followed by consolidation. It will depend
on the sort of parameters employed whether undrained or drained.
Figure 7.11 illustrates the total settlement, along section A-A, obtained
from PLAXIS. We can see that the maximum settlement is about 3 m between
the improved zones.
A good agreement is observed between the results coming from both meth-
ods which suggests that a reasonable estimation of the equivalent Youngs
modulus has been adopted.
7.5 Conclusions
It has been shown that although the starting value of undrained strength is
quite low, it is still possible to carry out a safe construction splitting the con-
struction in two main phases with a waiting period in between and introducing
geotextile reinforcement. The reinforcement is required only to cope with the
rapid construction of phase 1 but as soon as the foundation soil is allowed
to consolidate and gain strength, the role of the geotextiles becomes a lot less
important. In fact, stability analysis showed that the long term stability is
assured even without geotextile reinforcement.
Conclusions 89
A maximum consolidation settlement of the order of 3 m can be expected
between the improved zones. This settlement is expected to develop gradually
in a long period of time. Judging from the consolidation predictions, a fast
progress can be expected during the rst couple of years. This predictions
were conrmed later with measurements (see section 5.5).
Liquefaction risk is known to be an issue for hydraulically placed lls.
However, given the low earthquake activity in Belgium, such risk was shown
to be low. Safety factors against liquefaction were initially evaluated making
use of expected values of cone pressure in the embankment sand and they
were conrmed later with actual CPT data (see section 4.7).
8
Ground improvement by deep mixing
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
P. Menge & M. Van den Broeck
DEME, Belgium
8.1 Introduction
In recent years, an increased interest has been clearly demonstrated for design
and construction on low bearing capacity soils, principally in low lands where
many construction projects are conducted on soft alluvial clays, land reclaimed
with dredged materials, highly organic soils, etc. In order to cope with such
dicult subsoil conditions, various types of ground improvement methods
have largely evolved in the last decades as described by Van Impe W.F. (1989,
1997).
The deep mixing (DM) method originated in the early 1970s in the Scan-
dinavian countries and in Japan, almost simultaneously. The method could
be classied as a permanent soil improvement technique with addition of ce-
menting agents. Nowadays, binders such as cement, quicklime, y ash, blast
furnace slag, etc. are commonly employed to enhance mechanical and/or envi-
ronmental properties of the natural soil. The cementing agents can be injected
and mixed in place in either slurry or dry form. Furthermore, the mixing pro-
cedure can be purely mechanical or high-energy pressurized, both typically
making use of rotating mixing tools.
The deep mixing method is currently of great interest and is often applied
in near shore conditions for a number of applications such as improving bear-
ing capacity, reducing settlements, reinforcing slopes, earthquake mitigation,
etc. Worldwide statistics show the yearly growing importance of this technique
(Rathmayer, 1996; CDIT, 2002).
In Belgium, deep mixing techniques are still not largely implemented. Nev-
ertheless, scientic backed up experiences do shows the benets that this
ground improvement technique could bring in. In this chapter, local experi-
ences on land and underwater (near shore) will be described. The on land
experience makes use of dry deep mixing featuring mechanical mixing by
means of a blade. The underwater near shore experience deals with the im-
provement of the underwater embankment foundation soil. It features wet
deep pressurized intensive mixing.
92 Ground improvement by deep mixing
Blocks Single
Panels Grid
Figure 8.1. Types of application
8.2 Deep mixing applications
Worldwide DM methods have been implemented in special construction tech-
niques for onshore and oshore civil works. They have also been used for
improvement of foundations of a number of structures such as dams, em-
bankments, tanks, bridges, retaining structures, high-rise buildings, etc. On
land, the method has been employed on temporary stability works for deep
excavations, protection of adjacent structures and stabilization of slopes.
Deep mixing can be executed in mass or column stabilization and both
can be applied in many dierent ways. The soil can be stabilized either by
forming columns of stabilized soil (so-called column stabilization) or by sta-
bilizing the entire soil volume (so-called mass stabilization). Figure 8.1 gives
some examples of the conguration of columns and gure 8.2 suggests the
application for the combined mass and column stabilization.
Globally, local technologies are developed for new applications and for
specic geographic areas often by innovative contractors who are seeking
to develop their own variant of the method to match a particular project
challenge.
The large number of existing DM techniques have been classied by Bruce
(2001) on the following simple basis:
Is the cementious material injected wet (W) or dry (D)?
Is the binder mixed with the soil by means of rotary energy (R) only or is
the mixing enhanced by high-pressure jet (J) methods?
Is the mixing action only occurring near to the drilling tool (E), or is it
continued along the shaft (S) for a signicant distance above it, by way of
augers and/or paddles?
Figure 8.3 shows the classication proposed on these basis. Four categories
of methods have been identied: WRS, WRE, WJE and DRE.
Mechanism of stabilization 93
Embankment
Mass stabilization
Column stabilization
Figure 8.2. Combination of column and mass stabilization
Figure 8.3. Classication of deep mixing methods (adapted from Bruce, 2001)
8.3 Mechanism of stabilization
The stabilizing agents are in most of the cases, Portland cement and lime, but
also other binders have been more recently implemented. Some of these new
binders have been designed for clayey soils with high natural water contents or
94 Ground improvement by deep mixing
Figure 8.4. Lime stabilization mechanism (CDIT, 2002)
organic soils for which ordinary Portland cement or lime is not very eective.
A brief description of the mechanism of stabilization of lime and cement is
given in the following sections.
8.3.1 Stabilization with lime
When mixed with clayey soil, quicklime (CaO) reacts with pore water of
the soil to become slaked lime (CaO+H
2
OCa(OH)
2
). This reaction takes
places quickly and releases a large amount of heat. This brings in a reduction of
the natural water content of the soil which already represent an improvement
on its shear strength.
In presence of sucient water the hydrated lime dissolves into Ca
2+
and
OH

. Then, Ca
2+
ions exchange with cations on the surface of the clay miner-
als. The cation exchange reaction alters the characteristics of the water lms
on the clay minerals. In general the plastic limit of the soil is increased, reduc-
ing the plasticity index. Furthermore, under a high concentration of hydroxyl
ions (high pH), silica and/or aluminum in the clay minerals dissolve into the
pore water and react with calcium to form a water insoluble gel of calcium-
silicate or calcium-aluminate. This pozzolanic reaction goes on as long as the
high pH condition is maintained and calcium exists in excess. Figure 8.4 illus-
trates the lime stabilization mechanism in which the product of the pozzolanic
reaction cements the clay particles together.
8.3.2 Stabilization with cement-like binders
The most commonly used cement types for stabilization are Portland cement
and Blast furnace cement. Portland cements are inorganic binders obtained
Mechanism of stabilization 95
Cement slurry
Cluster of clay particles Hardened soil portions
Hardened cement portions
a) Immediately after mixing b) Long after hardening
Figure 8.5. Cement stabilization mechanism (CDIT, 2002)
by grinding to a high neness, Portland clinker alone, or most commonly in
combination with calcium sulfate (gypsum) acting as a set regulator.
In ordinary Portland clinker, tricalcium silicate (C
3
S) is the most abundant
phase present in amounts between about 50% and 70%. Dicalcium silicate
(C
2
S) usually constitutes 1530% of the clinker. Typical amounts of tricalcium
aluminate (C
3
A) are 510% and of the ferrite phase (C
4
AF) 515%. During
the hydration of the cement a C-S-H phase is formed and Ca(OH)
2
is released.
The rst hydration product has high strength which increases as it ages,
while Ca(OH)
2
contributes to the pozzolanic reaction as in the case of lime
stabilisation.
Figure 8.5 illustrates the cement stabilization mechanism. Immediately
after mixing it is possible to identify clay clusters and cement paste as separate
phases. Next, the strength of the stabilized soil will gradually increase due to
pozzolanic reactions within the clay clusters and hardening of the cement
paste.
Blast furnace cement is a mix of Portland cement and blast furnace slag
and shows a similar stabilization mechanism. Finely powdered slag does not
react with water but it has the potential to produce pozzolanic reaction prod-
ucts under high alkaline conditions. The SiO
2
and Al
2
O
3
contained in the
slag are actively released by the stimulus of the large quantities of Ca
2+
and
SO
2
4
from the cement, so that a hydration product is formed for which the
long-term strength is enhanced.
The complicated mechanism of stabilization has been simplied by Saitoh
et al. (1985) in gure 8.6 for the chemical reactions between clay, pore water,
cement and slag.
96 Ground improvement by deep mixing
Figure 8.6. Chemical reaction among clay, cement, slag and water (Saitoh et al.,
1985)
8.4 Methods of installation
The installation of columns stabilized with deep mixing methods requires the
use of specially designed machinery which basically consists of:
A binder feeding unit
A soil mixing machine for injection of binder into the ground
Figure 8.7 illustrates a scheme of a typical DM column installation
machinery. The binder feeding unit comprises various devices that measure
the quantity of the ingredients in the admixture and transport it to the soil
mixing machine. The plant generally include silos, automatic batching scales
and a slurry or air pump.
A variety of soil mixing tools have been manufactured for various im-
provement purposes, ground conditions and special applications. In general,
two categories of mixing tools can be identied:
Blade-based mixing tools: These tools have wide blades for excavation, in ad-
dition to paddles and/or short blades for cutting and mixing. The mix-
ing process is mainly carried out at the tip (or close to the tip) of the
tool (e.g. Fig. 8.8). The blades have a variety of shapes, dimensions and
orientations.
Belgian experience on on-land deep mixing 97
Binder A Binder B
Binder feeding
unit
Soil mixing
machine
Figure 8.7. Scheme of equipment for DM column installation
Figure 8.8. A typical mixing tool for the dry mixing method in Europe
Auger-based mixing tools: These tools have discontinuous or continuous he-
lical augers for drilling, in addition to paddles and/or short blades for
cutting and mixing. The mixing process is carried out in portions along
the drilling shaft.
8.5 Belgian experience on on-land deep mixing
The aim of this section is to illustrate the Belgian experience with the dry deep
mixing technique that could be implemented on, for example, improvement
of an embankment foundation on land.
A broad research project devoted to investigate the performance of deep
mixing methods for improvement of Flemish alluvial soils was carried out at
Ghent University (Ver astegui et al., 2005).
98 Ground improvement by deep mixing
0
2
4
6
8
10
0 2 4 6 8 10
D
e
p
t
h

(
m
)
qc
OCR
qc (Mpa) OCR
Columns
foundation level
Silty clay
GWT
Sand
Highly
organic
silty clay
(peat)
Tertiary
clayey
sand
Figure 8.9. Natural soil prole at the test site
The research was focused on the improvement, on land, of soft clay and
peat (often encountered in Flanders) with addition of cementing agents by
means of dry mixing. Lime and dierent types of cement (i.e. Portland cement,
blast furnace cement and others) have been employed. In the laboratory, the
improvement of small cylindrical laboratory-mixed specimens has been fol-
lowed up for a certain period of time. It has been found that blast furnace
cements work quite well for both soil types. In the eld, instrumented trial
embankments built on improved and non-improved ground showed the ben-
et of dierent binder dosages when lime-cement columns are installed. The
actual improvement in the testing site was assessed by eld and laboratory
testing.
8.5.1 Properties of untreated soils on land
The soil prole at the test site for the implementation of L-C column improve-
ment was dened after an extensive eld and laboratory testing campaign in-
cluding piezocone penetration tests, vane tests, dilatometer tests and borings
for sampling of disturbed and undisturbed specimens.
The CPT soundings showed the presence of soft alluvial soil in the upper
8 m of the prole overlying a clayey sand formation (Tertiary). Moreover,
the soft layer was not homogeneous but it consisted of two main sublayers,
corresponding to silty clay with sandy seams (from 0 to about 4 m below
ground surface) and a highly organic silty clay (from 4 to about 8 m below
the ground surface) with a sand content increasing with depth. Figure 8.9
Belgian experience on on-land deep mixing 99
Table 8.1. Physical properties of the natural soil
Index Silty clay Peat
Liquid limit 65.5 241.8
Plastic limit 22.8 135.0
Plasticity index 42.7 106.8
Natural water content 45.0 240.0
Organic content 1.7 1830
Natural carbonates content 5.1 11.2
Sand fraction, % 28.3 29.0
Wet density, g/cm
3
1.7 1.2
shows a typical CPT prole and the OCR estimated by DMT. The highly
organic silty clay is denominated peat for simplicity.
Disturbed samples taken at several depths from the silty clay and highly
organic silty clay (peat) have been tested on physical properties. Some pa-
rameters of each soil type are summarized in table 8.1.
The undrained shear strength of the silty clay and the peat at the test-
ing site has been determined by means of CPTU soundings, eld vane tests,
dilatometer tests and triaxial testing. Figure 8.10 summarizes all measure-
ments. The undrained strength prole in the gure shows that c
u
ranges, in
general, from 20 to 40 kPa. The lowest values do obviously correspond to the
peat.
0
1
2
3
4
5
6
7
8
0 20 40 60 80 100 120
c
u
(kPa)
D
e
p
t
h

(
m
)
CPTU
Dilatometer
Triaxial UU
Field Vane 1 (peak)
Field Vane 1 (resid)
Lab vane (peak)
Silty clay
Sand
Highly organic
silty clay (peat)
Figure 8.10. Undrained shear strength prole of natural soils at the testing site on
land
100 Ground improvement by deep mixing
8.5.2 Binders employed on land
In this research on dry mixing on land, quicklime and cement have been
chosen as binders; however, dierent types of cement have been tried out
(i.e. Portland cement, CEM I, composite cement, CEM II, and Blast furnace
cement, CEM III). The name listing employed here complies with the standard
EN 197-1; for example, CEM I 42.5 refers to a Portland cement with a nominal
compressive strength of 42.5 MPa.
Lime and cement have been employed in dierent proportions (e.g. L/C
50/50, 20/80, 0/100, percentages in weight). Dry mixing was implemented in
the laboratory as well.
The quantity of binder has been set to a range varying from 100 to 200
kg/m
3
(kg of binder per m
3
of natural soil). CEM I, CEM II and CEM III
have been employed for the stabilization of silty clay. On the other hand, CEM
II and CEM III have been used in case of the peat.
8.5.3 Lime-cement stabilization in the laboratory
[Link] Preparation of specimens
The natural soil samples collected from a number of borings have been rst
selected and then thoroughly homogenized prior to stabilization in the labora-
tory. A dough mixer has been employed in the laboratory for mixing of natural
soil at the natural water content with the binders (added dry). A mixing time
of about 5 minutes was implemented. Immediately afterward, small specimens
(H =9 cm, =4.5 cm) have been molded either by static compaction (for silty
clay) or by pouring (for peat) into plastic split cylindrical molds.
The stabilized specimen have been sealed with paran and stored under
water in a conditioned room at 20

C.
[Link] Unconfined compression tests
Unconned compression (UC) tests have been carried out at specic time
intervals up to 90 days after the preparation of the stabilized specimens
(Ver astegui et al., 2004). Some results have been summarized in gures 8.11
and 8.12. Figure 8.11 shows that the combination L/C-20/80 with blast fur-
nace cement leads to the highest UC strength (UCS) for stabilized silty clay.
In fact, a ratio UCS
stab
/UCS
natural
40 has been reached in 60 days with
a dosage of 150 kg/m
3
; moreover, UCS seems to still increase.
On the other hand, the combination L/C-20/80 with Portland cement
shows little extra improvement after the rst month; nevertheless, the ratio
UCS
stab
/UCS
natural
reaches a value of the order of 12. The composite binders
with CEM II/B do show that, the higher the ratio of quicklime/cement the
smaller the UC strength; however, quicklime plays a very important role on
Belgian experience on on-land deep mixing 101
Silty clay
0
500
1000
1500
2000
2500
3000
0 50 100 150
Time (days)
U
C
S

(
k
P
a
)
L/C-20/80 (150)
CEM I 42.5
L/C-0/100 (150)
CEM II/B 32.5
L/C-20/80 (150)
CEM II/B 32.5
L/C-50/50 (150)
CEM II/B 32.5
L/C-20/80 (150)
CEM III/B 42.5
Figure 8.11. Unconned compressive strength of silty clay
Highly organic silty clay (peat)
0
20
40
60
80
100
120
140
0 50 100 150
Time (days)
U
C
S

(
k
P
a
)
L/C-0/100 (150)
CEM II/B 32.5
L/C-20/80 (150)
CEM II/B 32.5
L/C-0/100 (150)
CEM III/A 32.5
Figure 8.12. Unconned compressive strength of peat
the quality of the mix, as the scatter of UCS decreases with an increasing
amount of lime.
Figure 8.12 illustrates the development of the UC strength of stabilized
peat with time. Clearly, the benet of the lime here was less signicant for the
strength and mix quality. The UCS improvement on samples stabilized with
CEM II/B seems to cease after 1 month, while specimens mixed with blast
furnace cement, CEM III/A, show a slow but continuous increment. A ratio
102 Ground improvement by deep mixing
UCS
stab
/UCS
natural
ranging from 2 to 3 has been evaluated after 90 days.
Note that the specimens have not been subjected to any overburden in the
curing stage.
8.5.4 Lime-cement stabilization in-situ
To the extent of controlling the quality of the DM method itself in the eld, a
number of trial stabilized columns, =0.6 m, were installed on the on land test
site with the dry mixing technique (more details were reported by Ver astegui
et al., 2004).
In the installation phase, the dry composite binder has been injected, by
means of compressed air, at pressures not higher than 5 bar through a tubing
down to the mixing tool. The DM column is formed below the mixing tool
lifting the mixing auger while rotating continuously.
[Link] Quality control of lime-cement trial columns
Four trial columns were initially installed in the site on land. A scheme
summarizing the characteristics and composition of each column is given in
gure 8.13.
Column 1 has been stabilized with 85 kg/m
3
of unslaked lime, column 2
with 130 kg/m
3
of cement (CEM II/B-M 32.5), column 3 with 130 kg/m
3
of a
blend (50/50) of unslaked lime and cement (CEM II/B-M 32.5), and column
4 with 170 kg/m
3
of unslaked lime.
Within the framework of quality control of stabilized columns, the extrac-
tion of the whole column would allow for a good evaluation of the homogeneity
Tertiary formation
Column 1
L/C-100/0
85 kg/m
3
Column 2
L/C-0/100
130 kg/m
3
Column 3
L/C - 50/50
130 kg/m
3
Column 4
L/C - 100/0
170 kg/m
3
0.6m
8 to 9m
Heterogeneous
soft soil deposit
from the
Quaternary
Figure 8.13. Layout of trial columns 1, 2, 3 and 4
Belgian experience on on-land deep mixing 103
Figure 8.14. Partial excavation of trial columns (a) No. 1 mixed with 85 kg/m
3
of
L/C-100/0 (b) No. 2 mixed with 130 kg/m
3
of L/C-0/100 (c) No. 3
mixed with 130 kg/m
3
of L/C-50/50 and (d) No. 4 mixed with 170
kg/m
3
of L/C-100/0; all implementing CEM II/B-M 32.5
of the mix; however, such practice has proved to be expensive and dicult
to put in practice. For that reason an alternative way has been chosen here;
that is, partial excavation (to about 5 m below the ground level) and vertical
borings through the entire column.
Columns 1 to 4, stabilized with dierent composite binders, have been
excavated for visual inspection (Fig. 8.14). The inspection of the columns has
104 Ground improvement by deep mixing
UC tests on L/C column specimens (~65 days)
0
50
100
150
200
250
300
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
.

s
t
r
e
n
g
t
h

(
k
P
a
)
qu [kPa] 95.9 42.6 69.7 51.2 172.0 75.9 104.3 93.4
Silty
clay
Peat
Silty
clay
Peat
Silty
clay
Peat
Silty
clay
Peat
L/C-100/0
85 kg/m
3
Col. 1
L/C-0/100
130 kg/m
3
Col.2
L/C-100/0
170 kg/m
3
Col. 4
L/C-50/50
130 kg/m
3
Col.3
Strength expected
from laboratory
tests for Col. 3
Figure 8.15. UC strength of specimens from the eld inspection
taken place about 2 months after the installation. The following remarks were
been made:
Column 1 (mixed with L/C-100/0, 85 kg/m
3
) shows a rather uniform shaft
diameter. Even though the soil in the neighborhood is very plastic, the visual
inspection shows that the binder has been properly mixed.
Column 2 (mixed with L/C-0/100, 130 kg/m
3
) shows some discrepancy
on its diameter. It has also been noticed that the quality of the mix is not as
good as in column 1 especially where the plasticity of the soil is high (little
grains of hardened cement have been found there).
Column 3 (mixed with L/C-50/50, 130 kg/m
3
) shows a very uniform,
homogeneous and well shaped shaft. The quality of the mix seems to be quite
good along the exposed portion of the column. The binder employed in this
column is a blend of lime and cement.
Column 4 (mixed with L/C-100/0, 170 kg/m
3
) shows a uniform shaft
diameter as well. The binder seems to be very well mixed, as it was the case
for column 1. The dosage employed here was rather high, still, no sign of
binder spreading (outside the column) has been found.
Just after the inspection, a few specimens have been sampled (by means of
horizontally pushed-in thin wall tubes) from the exposed section of the columns.
Figure 8.15 summarizes the average results of UC tests on these specimens.
The results are consistent with the visual inspection. Column 2, in which
the highest strength was expected, does not show a good performance. Indeed,
the visual inspection had come across the fact that the mixing quality of Col-
umn 2 was the poorest. It has been suggested here that mechanical mixing
of cement in a plastic soil is a very dicult task. However, when a blend
lime-cement is employed, the lime reduces the plasticity of the soil facilitating
the homogenization of the stabilized mass, which is later reected in a higher
strength.
Belgian experience on on-land deep mixing 105
0
1
2
3
4
5
6
7
8
9
10
0 50 100 150 200
Carbonate content [%]
D
e
p
t
h

(
m
)
Natural content
After stabilisation
Figure 8.16. Distribution of binder along column 4, installed with the method
implemented here, in terms of carbonate content
Moreover, gure 8.15 illustrates the strength expected from laboratory
tests for column 3. It seems that the laboratory tests have over predicted the
strength in the eld with a ratio UCS
lab
/UCS
field
less than 2. Ratios ranging
from 2 to 5 are usually reported in literature for dry mixing methods.
As there were no straightforward means of measuring the amount of
cement in the soil due to its complex chemistry, it was decided to trace the
amount of lime by means of a simple standard physical test. The test provides
the amount of carbonates in the soil making use of a remotion agent (HCl).
The results of the tests on specimens from the trial column 4 (that is mixed
with lime only) are shown in gure 8.16.
Judging for the carbonate content of specimens from the column as com-
pared to the natural lime content of the soil in that area, one could conclude
that a rather uniform distribution of the binder has been achieved during the
installation. This might also be the case for the other columns where compos-
ite binders were employed.
[Link] Trial embankments for testing the performance
of improved soils
From the laboratory research outcome for the soils at the site on land, it was
decided to use a combination quicklime/blast furnace cement (CEM III/B
42.5) at L/C-20/80 for silty clay and at L/C-0/100 for peat.
106 Ground improvement by deep mixing
CL
Fill
Silty
clay
Peat
L/C-20/80
L/C-0/100
7.5 m
2 m
4 m
4 m
A
Silty clay
0 kg/m
3
15 m
Peat
0 kg/m
3
60 m
B
Silty clay
100 kg/m
3
Peat
200 kg/m
3
C
Silty clay
150 kg/m
3
Peat
200 kg/m
3
D
Silty clay
200 kg/m
3
Peat
200 kg/m
3
Cross section
Planview
and binder
dosage
Figure 8.17. Layout of trial embankments
Next to the trial columns described in the previous section, 3 trial DM
zones were built to evaluate the behavior of the foundation soil when subjected
to an embankment load. Also a reference untreated foundation soil was loaded.
The aim of these loading tests was to study the response of columns improved
with dierent binder dosages.
Figure 8.17 illustrates the general layout of each embankment and the
dosage per specic soil layer. Note that the dosage for the peaty layer was
xed to 200 kg/m
3
in all zones. The spacing (axis to axis) between columns
was set to 1.8 m in a triangular arrangement. In each zone, the embankment
ll aimed at a net surcharge of about 30 kPa.
Figure 8.18 illustrates the results of the assessment of column installation
eects on the soil nearby a stabilized column within the column array for trial
embankments. To that aim, dilatometer tests were performed in the close
vicinity of a DM column before and 2 months after installation. It was clear
that when comparing the state of the natural soil before and after installation
in terms of the constrained modulus (correlated from DMT measurements)
the installation eects seemed to be not detrimental at all. On the contrary,
the modulus of the soil shows an increase, where more sandy soil is present.
No stress relaxation was observed around the column due to the combined
action of the mixing tool and the compressed air.
The in-situ evaluation of the improvement of this stabilized columns was
carried out by CPT tests (5 months after installation). The CPT was per-
formed through the axis of the column. Figure 8.19 illustrates the CPT proles
in the natural soil (untreated) and in the column axis. Clearly, a remarkable
improvement, in terms of cone penetration pressure, can be observed in the
upper silty clay layer where the ratio of q
c
column/q
c
natural increases with
depth to values of the order of 30 to 40. Similarly, an important improvement
has been evaluated in the peaty layer with q
c
column/q
c
natural ranging from
4 up to 7. The more sandy zones are clearly identied by the peaks of q
c
.
Belgian experience on on-land deep mixing 107
0
1
2
3
4
5
6
7
8
0 50 100 150
Constrained modulus (MPa)
D
e
p
t
h

(
m
)
Before installation
After installation
(2 months later)
Column
0.3 m
Dilatometer
membrane
Silty clay
Sand
Highly organic
silty clay (peat)
Figure 8.18. Evaluation of installation eects on adjacent soil within the trial
embankment columns
Figure 8.20 shows the outcome of the settlement monitoring of the trial
embankments by means of settlement tell tale plates. A period of about 2
months was set between the end of column installation and the initiation of
trial embankments construction.
As expected, the reference embankment (A) shows the largest settlements
and a very rational tendency was observed for the trial embankments on
improved soil (B, C and D). The binder dosage for embankment B seems to
be not high enough to allow for some signicant benet.
Also lateral deformations have been monitored by means of inclinome-
ters installed on one side of the trial embankments (at about 1 m away from
the side boundary). Figure 8.21 reects these measurements, 1 month af-
ter loading. As expected, the horizontal deformations in a vertical close to
embankment D are the smallest as compared to the values for embankment
C. Embankment B, on the other hand, shows to induce by far larger lateral
deformations, conrming the indication that such binder dosage employed in
the soil deposit is insucient to allow for a considerable improvement.
8.5.5 Remarks on the experience of dry deep mixing on land
The laboratory research for the application of deep mixing on land has shown
good potential for the stabilization of silty clay and peat by blast furnace
108 Ground improvement by deep mixing
0
1
2
3
4
5
6
7
8
0 5 10 15 20
q
c
(MPa)
D
e
p
t
h

(
m
)
In the column axis
In the natural soil
Performed 5 months after
column installation
Silty clay
Sand
Highly organic
silty clay (peat)
Tertiary clayey sand
Figure 8.19. Evaluation of the improvement in situ by means of CPT on a column
from embankment C
-0.18
-0.16
-0.14
-0.12
-0.1
-0.08
-0.06
-0.04
-0.02
0
0 10 5 15 20 25 30
Time (days)
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
)
Embankment D
Embankment C
Embankment B
Embankment A
Embankment
Figure 8.20. Settlement of the trial embankments
cements. A ratio UCS
stab
/UCS
natural
40 has been reached for silty clay
(in 60 days) with L/C-20/80 (150 kg/m
3
). The tests on peat indicate a slow
but continuous improvement with a ratio UCS
stab
/UCS
natural
ranging from
2 to 3, after 90 days.
Deep mixing assessment on the underwater site 109
0
1
2
3
4
5
6
7
8
9
-5 0 5 10 15 20
Horizontal deformations (mm)
D
e
p
t
h

(
m
)
Embankment D
Embankment C
Embankment B
Measurements 1
month after
loading
Level of the fill
overburden
Figure 8.21. Settlement of the trial embankments
In the eld, the assessment of the DM column installation eects (using
this type of dry mixing method) with the column array for trial embankments
allows to conclude for no stress relaxation. The evaluation of the improvement
by means of CPT (performed 5 months after installation) shows a remarkable
improvement in the silty clay layer where q
c
column/q
c
natural increases with
depth to values of the order of 30 to 40. In the peaty layer, on the other hand,
q
c
column/q
c
natural ranges from 4 up to 7. Moreover, from a quality control
it was observed that the strength of specimens stabilized in the laboratory is
greater than the strength achieved in the eld by a factor less than 2.
The monitoring of trial embankments (on land) aimed at nding out the
level of the benet of the columns on the settlements and lateral deformations.
A settlement reduction of about 65% was evaluated at the highest binder
dosage implemented (200 kg/m
3
). The lowest binder dosage of 100 kg/m
3
was
found insucient to produce considerable improvement, at least on the soil
conditions studied here.
8.6 Deep mixing assessment on the underwater site
The soil investigated here, as described in chapter 6, is a soft deposit of ne
grained material, result of a prolonged sedimentation and self-weight consol-
idation process of dregs removed from the waterways within the harbor of
Antwerp. In many harbor areas all around the world, there is an increasing
need of reclaimed land. This fact has encouraged the design and ongoing con-
struction in the Antwerp harbor of a partially submerged 27-m high sand
embankment on the soft material previously mentioned. Obviously, the pres-
ence of such soft foundation layer caused concern for the overall stability;
therefore partial improvement of the material by deep mixing was proposed.
110 Ground improvement by deep mixing
This section focuses on the laboratory and eld investigation carried out
for the evaluation of the improvement of the soft ne grained soil with cement.
Initially, the eect of several types of cement was studied in the laboratory.
From those results a blast furnace cement was chosen as most suitable for the
application in the eld. Finally, a eld inspection was carried out to asses the
actual improvement of the deep mixing columns installed by the SSI technique
(SSI is a technique patented by HSS, Dredging International-Belgium).
8.6.1 Properties of the articially cemented soil in the laboratory
[Link] Preparation and mixing of specimen
The soil collected from the soft deposit was thoroughly homogenised and
remolded prior to mixing with cementing agents.
A dough mixer was employed here to mix the soil and a slurry of cement.
The dosage of binder for mixing with soil was set to 275 kg/m
3
, the water/
cement ratio of the slurry was set to 0.8 and a mixing time of about 10 minutes
was implemented. This extended mixing time was meant to allow for more
intensive mixing; however, only a slight dierence in strength was observed
when compared to specimens mixed for 5 minutes (less than 5% after 7 days).
Cylindrical specimen with a diameter of 57 mm and a height of 115 mm
were prepared by pouring the mix into split plastic moulds. The moulds were
later sealed with paran lm and stored under water in a conditioned room
at 10

C with no overburden whatsoever acting on the specimen. In addition,

some specimen were cured under water at 20

C in order to study the eect of

the temperature on the development of the improvement.
[Link] Binders
At the initial stage of this project, a number of dierent types of cement have
been employed in the laboratory. A short description (according to EN 197-1)
of the binders is given below:
Binders A, B, and C are all blast furnace cements, CEM III. Binder C
has the greatest blast furnace slag content (CEM III/B). Binders B and
C classify at a nominal strength of 42.5 MPa while binder A has only 32.5
MPa.
Binder D is a Portland cement, CEM I, with a nominal strength of 52.5
MPa.
Binder E is a commercially available binder specically designed for sta-
bilization of soil.
Binder F is a cement typically used for soil grouting purposes.
[Link] Compressive strength
A large number of unconned compression tests have been performed at
several time intervals (i.e. 7, 14, 28, 56, 84, 120, 240 and 550 days). The
Deep mixing assessment on the underwater site 111
0
500
1000
1500
2000
2500
3000
0 100 200 300 400 500 600
Time (days)
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

(
k
P
a
)
Binder A
Binder B
Binder C
Binder D
Binder E Binder F
Binder E
Binder F
Blast furnace
cement
CEM III 42.5
Blast furnace
cement
CEM III 32.5
Portland cement
CEM I 52.5
Figure 8.22. UCS of cement stabilised specimen cured underwater at 10

C
results of the testing programme on specimen cured under water at 10

C
have already been reported by Van Impe et al. (2004) and are summarized in
gure 8.22.
From the group of binders tested here, it seems that the blast furnace
cements (binders A, B and C) perform quite well, showing a continuous in-
crease of the UC strength. Binders B and C (both CEM III 42.5) do show
an unconned compressive strength of the order of UCS 2.2 MPa after 550
days. The Portland cement (binder D), on the other hand, allows for more
rapid hardening in the rst days. In fact, it shows the highest UC strength
during the rst month. However, the improvement provided by Portland ce-
ment seems to decline afterwards for some period to nally pick up again after
some 3 months. The understanding of why systematically this interval of
the interplay cement-soil occurs is subject to further research today. Anyhow,
the nal compressive strength of Portland cement remains lower than that
given by the blast furnace cements B and C. The other binders (E and F)
seem to produce little improvement for such high dosage (UCS < 0.7 MPa
after 550 days).
The strain at failure of specimen cured under water at 10

C, illustrated
in gure 8.23, was measured externally (from top to bottom cap of a triaxial
cell) by LVDT. The gure provides some information about the ductility of
the stabilised mass. In spite of some scatter it seems possible to establish a
general tendency of behaviour for each binder mix. Overall, the strain at fail-
ure (ranging from 0.9% to 4%) decrease rapidly with increasing UC strength.
The brittleness increases obviously with increasing UCS values. From the
112 Ground improvement by deep mixing
0
500
1000
1500
2000
2500
3000
0 0.01 0.02 0.03 0.04 0.05
Vertical strain at failure
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

(
k
P
a
)
Binder A Binder B
Binder C Binder D
Binder E Binder F
Portland
cement
Blast furnace
cement
Ageing
Figure 8.23. Strain at failure of cement stabilized specimen cured under water at
10

C
results it can also be deduced that, as ageing increases, specimen mixed with
Portland cement tend to yield at smaller axial strains than specimen mixed
with blast furnace cement, even tough the strength of specimen mixed with
Portland cement is considerably lower.
[Link] Stiffness modulus
Measurements of small-strain modulus were also performed by means of ben-
der element testing at dierent time intervals for some specimen (cured under
water at 10

C) mixed with blast furnace cement (binder C) and with Portland

cement (binder D) only.
The bender element test set up employed here is given in gure 8.24.
The principle of this non-destructive method is simple and well know from
literature (Dyvik and Madshus, 1985).
As an example, gure 8.25 illustrates the S-wave arrival time measured for
specimen stabilised with blast furnace cement at several curing time intervals
using an input sinusoidal pulse with a frequency of 4 kHz. Each specimen was
tested for unconned compression to measure UCS. As expected, a rather
linear relationship between G
0
, E
0
and UCS is observed.
Figure 8.26 summarizes the Youngs modulus at small strain E
0
evalu-
ated here for specimen mixed with Portland and blast furnace cement. The
modulus for the Portland cement was found to be slightly higher but still, a
single linear correlation has been proposed for both cements: E
0
714 UCS.
Similarly, gure 8.26 illustrates the secant Youngs modulus evaluated from
Deep mixing assessment on the underwater site 113
SIGNAL ANALIZER
t
FUNCTION GENERATOR
Shape, freq, amp
TOP
CAP
BOTTOM
CAP
SOIL
SPECIMEN
TRANSMITTER
ELEMENT
RECEIVER
ELEMENT
Figure 8.24. Bender elements testing setup
0 200 400 600 800 1000 1200 1400
S-wave arrival time (s)
7 days
14 days
28 days
42 days
98 days
193 days
Curing
period
Figure 8.25. Shear wave arrival time measured at several curing time intervals on
specimen mixed with blast furnace cement and cured under water at
10

C
unconned compression tests. Even if trend shows some scatter, the data
could be more or less linearly correlated to UCS as well. It has been es-
timated as E
s50
110 UCS. This trend is considerably low when com-
pared to the Japanese experiences reported by Saitoh et al. (1985) where
114 Ground improvement by deep mixing
0
100
200
300
400
500
600
700
0 200 400 600 800 1000 1200
UCS (kPa)
Y
o
u
n
g
'
s

m
o
d
u
l
u
s

(
M
P
a
)
Portland cement
Blast furnace cement
from BE testing
from UC testing
Eo = 714 x UCS
Es
50
= 110 x UCS
7 days
98 days
42 days
14 days
173 days
78 days
7 days
15 days
28 days
49 days
120 days

f
=1.8%
120 days

f
=0.9%
28 days

f
=1.2%
28 days

f
=1.3%
Figure 8.26. Youngs modulus at small strain levels (E
0
) and secant Youngs mod-
ulus at 50% of deviatoric stress (E
s50
) versus UCS
350 UCS < E
s50
< 1000 UCS; however, it falls within the range of many
other correlations proposed worldwide in the literature (Porbaha et al., 2000).
Overall, the modulus of the Portland cement is slightly higher than that
given by the blast furnace cement. In general, E
0
remains about 7 times E
s50
.
[Link] Effect of the temperature
In an attempt to more reliably recreate the conditions in the eld, a large
cylindrical specimen with a height H 0.8 m and diameter 0.6 m was
prepared in the laboratory employing blast furnace cement, with the aim of
evaluating and monitoring the temperature changes due to exothermic reac-
tions within the stabilized mass.
The virgin soil was kept at a temperature of 10

C prior to mixing. After

mixing of the soil and blast furnace cement slurry in a concrete mixer, the
stabilized mass was poured into a large plastic mold (also stored at 10

C and
with the above mentioned dimensions) where eight temperature transducers
(labeled T1, T2, ... , T8) were installed at dierent locations within the sample.
A few small cylindrical specimen were also prepared and cured (under
water at 10

C) following the ordinary procedure described in a previous sub-

section.
The temperature measurements within the stabilized mass over a period of
56 days are illustrated in gure 8.27. The readings of all temperature transduc-
ers do show a common trend. Immediately after mixing a sudden temperature
increase was observed. After 3 days a maximum temperature of about 25

C
Deep mixing assessment on the underwater site 115
3 days
0
5
10
15
20
25
30
0 10 20 30 40 50 60
Curing time (days)
T
e
m
p
e
r
a
t
u
r
e

(

C
)
T1 T2 T3 T4
T5 T6 T7 T8
T8
T1
T5 T6 T7
T2 T3 T4
Conditioned room
temperature
Temperature in the
stabilised mass
0.60 m
0.80 m
Large specimen
Figure 8.27. Hydration temperature monitoring on large specimen mixed with
blast furnace cement and stored in a conditioned room at 10
o
C
0
200
400
600
800
1000
1200
1400
0 20 40 60 80 100 120
Time (days)
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

(
k
P
a
)
Small specimens from the routine lab
testing (10C)
Specimens from large specimen mix,
following routine procedure (10C)
Core specimens sampled from the
large specimen cured at 10C
Core specimens
(=50 mm) from
the large
specimen mixed
and casted in the
laboratory ( =
0.60 m)
Small specimens
from the routine
laboratory testing
( = 57 mm)
Figure 8.28. UCS of core specimen from a large stabilized specimen
was reached. Finally, the temperature in the large specimen seems to grad-
ually decrease; after 56 days, the temperature (about 11.7

C) leveled out at
values only slightly over the conditioned room temperature (10

C).
By the end of the temperature monitoring some core samples were taken
from the large specimen. Figure 8.28 shows the UCS of such core samples. The
116 Ground improvement by deep mixing
0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 50 100 150 200 250 300
Time (days)
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h

(
k
P
a
)
Blast furnace
cement (20C)
Blast furnace
cement (10C)
T=20C
T=10C
Figure 8.29. Eect of the curing temperature on the UCS of specimen mixed with
blast furnace cement
gure also indicates the UCS of small specimen from the routine laboratory
testing as described in a previous subsection. Clearly, the UCS of the large
specimen cores doubles the UCS values of the small specimen. This suggested
that the transient temperature increase due to the exothermic reactions within
the large specimen were imposing such notable dierence. Indeed, the larger
the sample, the slower the heat dissipation and so the higher the UCS to be
expected.
In order to study the eects of the curing temperature on the UC strength
of the stabilized dredged material an extra series of tests has been carried out;
this time on small specimen mixed with blast furnace cement, cured under
water at 20

C. The results (Fig. 8.29) demonstrated that the strength of the

samples stabilized with blast furnace cement is notoriously aected by the
temperature. The hydration of the blast furnace cement clearly benets from
high temperatures; in fact, the UCS of samples cured under water (up to 200
days) at 20

C is, at all times, about 1.7 to 2 times larger than the UCS of
specimen cured at 10

C.
8.6.2 Properties of the cemented soil in the eld
The experimentation for the evaluation of properties of the cemented soil in
the eld consisted of core sampling of specimen from trial columns to proceed
later on with unconned compression tests in the laboratory.
Deep mixing assessment on the underwater site 117
Figure 8.30. Mixing tool employed in the eld
[Link] Installation of trial columns
The trial deep mixing columns ( 1.9 m) were installed in the site
(underwater) with the SSI technique from a jack-up platform.
Only blast furnace cement (binder C) was used for the eld experimen-
tation. The cement was mixed with water an transformed into a slurry
(w/c =0.8) on land. The cement slurry was pumped to the jack-up platform
by means of oating pipes. In order to optimize the column installation rates
the jack-up platform was provided with a moon pool to allow the installation
of 22 to 24 columns in each zone covered by the platform. State of the art
positioning systems ensured a very precise location of each column.
The SSI technique makes use of pressurized mixing by means of a mixing
tool provided with 2 sets of nozzles distributed all along the full diameter of the
column (Fig. 8.30). The mixing tool is xed to a main drilling rod and each
set of nozzles is connected to independent injection systems (Fig. 8.31). A
high-pressure injection system (of the order of 20 to 30 MPa) cuts the soil
and allows for intense mixing while the low-pressure injection system (up to
5 MPa) adds the remaining amount of cement slurry to fulll the required
dosage. Injection of the cement slurry takes place during the downwards and
upwards operation of the drilling rod. A more detailed description of the instal-
lation and performance of the method can be found elsewhere (Van Mieghem
et al., 2004). All drilling and injection execution parameters were automati-
cally controlled to accomplish a binder dosage of 275 kg/m
3
approximately.
[Link] Evaluation of improvement in the field
A number of core specimen (=100 mm) obtained over the full length of
the trial deep mixing columns were tested to evaluate the actual unconned
compressive strength.
118 Ground improvement by deep mixing
Figure 8.31. SSI set up and methodology
Figure 8.32 compares the strength evaluated from core specimen to that
obtained from laboratory prepared specimen (after a curing period of 56 days).
The UCS in the eld ranges from 2 to 5 MPa in the upper 5 m and from 5
to 8 MPa in the lower zone, where a higher content of sand was observed.
While, starting from the laboratory investigation, a quite optimistic UCS
value remains below 0.9 MPa for similar conditions.
This discrepancy suggested that the ordinary practice of mechanical mix-
ing (with a dough mixer) of specimen in the laboratory severely underesti-
mates the actual strength of columns installed in the eld with pressurized
more intensive mixing procedures such as the SSI method (dierences up to
a factor of 2 to 5).
Deep mixing assessment on the underwater site 119
0
1
2
3
4
5
6
7
8
9
0 4 8 10
Unconfined compressive strength (MPa)
D
e
p
t
h

(
m
)
Strength of
samples mixed
in the laboratory
(also after a
curing period of
56 days)
Dredged
material
Sand
6 2
Figure 8.32. Unconned compressive strength of core specimen from SSI-stabilized
dredged material in the eld
8.6.3 Laboratory versus in-situ behavior
In order to evaluate the reasons for discrepancy between laboratory and eld
test results, a laboratory reconstituted and mechanically mixed specimen on
the one hand and undisturbed core samples from columns on site, on the
other hand, were analyzed by means of Scanning Electron Microscopy (SEM),
aiming at investigating the microstructure and composition of each specimen.
Figure 8.33 shows both specimen (4 cm4 cm1 cm) carefully cut with
a water-cooled sawing system starting from stabilized samples, either mixed
in the laboratory, or in the eld by the SSI technique. It is obvious, already
from this pictures, that the specimen dier in texture. Until the moment
of the microscopic analysis, the sample from the laboratory was about 300
days old and had been kept sealed, under water, in a T =10

C conditioned
room. On the other hand the sample from the eld was approximately 270
days old; this specimen was cored from a trial SSI column about 3 months
after its installation and then kept under water as well, until the day of SEM
analysis.
The presence of large pores in a considerable amount in the laboratory
specimen is evident. This is by far less pronounced in the SSI improved
eld specimen where a more compact and more homogeneous texture can
be observed. At this point it may be stated that the mixing in the labora-
tory (by means of a dough mixer) could have caused the incorporation of air
bubbles (large pores). Pores of smaller diameter observed in both specimen
probably are produced during the cement hydration process.
120 Ground improvement by deep mixing
(a)
(b)
Figure 8.33. Cement-stabilized specimen mixed by: (a) laboratory dough mixer,
(b) pressurized in-situ mixing tool
Figure 8.34 does illustrate the same specimen but, this time, with an
amplication factor of 1200. It is again quite clear that the micro structure
is diverse. The specimen from the eld has in general a much more homo-
geneous structure with a more regular distribution of hydration products,
such as the calcium silicate hydrate (C-S-H phase) and the calcium hydroxide
(CH). On the other hand, the mechanically mixed laboratory sample shows
a rather heterogeneous skeleton where the unaged morphology of the C-S-H
phase suggests still a lower degree of hydration.
Here it may also be suggested that the much more intensive high pressure
mixing in the eld did play an important role (the specic area around each
soil particle could be reached by the binder, by far better). It seems that
the high-pressure SSI mixing in the eld has improved the distribution of
cement around the soil particles and as a consequence a faster hydration
and hardening has been taking place. In the laboratory, where purely simple
Deep mixing assessment on the underwater site 121
(a)
C-S-H
(II)
C-S-H
(III)
CH
large
crystal
(b)
C-S-H
CH
Figure 8.34. SEM analysis: (a) Laboratory specimen (b) eld specimen at an am-
plication factor of 1200
mechanical mixing with a dough mixer was put into practice, the cement may
have not been so well distributed and is only reaching clusters of soil particles.
Finally, gure 8.35 illustrates the samples from the laboratory and the
eld with an amplication factor of 1700. Also here the same pattern was
observed; the structure of the eld specimen looks much more homogeneous
than the laboratory specimen. A matured C-S-H phase can be recognized
in the eld specimen together with uniformly distributed CH crystals that
cover almost completely the soil. On the other hand, the mechanically mixed
laboratory sample shows a much more heterogeneous composition including
also ettringite (AFt phase) that is formed during the early hydration process
(this phase is usually absent in matured and well hydrated cement pastes
(Odler, 2000)). Overall, judging on the morphology of the dierent cement
hydration products in the pictures, a by far less advanced degree of hydration
could be perceived in the laboratory prepared samples.
122 Ground improvement by deep mixing
(a)
AFt
AFt
CH
(b)
CH
C-S-H
Figure 8.35. SEM analysis: (a) Laboratory specimen (b) eld specimen at an
amplication factor of 1700
9
Construction and monitoring of embankment
W.F. Van Impe & R.D. Verastegui Flores
Laboratory of Geotechnics, Ghent University, Belgium
J. Van Mieghem & A. Baertsoen
Ministry of Flanders, Belgium
P. Menge & S. Vandycke
DEME, Belgium
9.1 Introduction
Today, the underwater embankment is still under construction. It has been
built in stages by spreading sand in layers of approximately 2 meters allowing
a period of time (1 to 2 months) in between. Today, a 70% of the total height
of the embankment has been reached and now a much longer period is being
allowed for consolidation.
Moreover, the quality of the embankment sand has been continuously con-
trolled by means of CPT executed at several stages during the construction.
With the aim of increasing the safety, geotextile reinforcement was in-
stalled in the embankment slope at the open dock side.
Already before the initiation of the construction, instrumentation was
mounted in the foundation layer to allow the monitoring of excess pore water
pressures (PWP) and displacements due to the embankment load that could
help to continuously check the behavior of the soil. Piezometers were installed
at 3 dierent levels within the soft soil layer at dierent locations. Similarly,
exible tubes were placed at 4 locations across the dock to monitor vertical
displacements by measuring water pressure changes (water height relative to
a reference) with a probe that is displaced inside the tube.
The outcome of the monitoring of pore water pressures and settlement
during the construction is given in the following sections. Also, the quality
control of the embankment sand is briey illustrated next.
9.2 Construction
The sand used for the underwater lling operation was mainly obtained from
the excavation works and dredging residues of an almost simultaneous con-
struction of a dock nearby. The sand was selected on the basis of its grain
size distribution and content of nes so that optimum results of density and
shear strength are obtained when it is hydraulically placed. Tests showed
124 Construction and monitoring of embankment
Figure 9.1. Layout of geotextiles and geocontainers
that the method of placing the sand could yield a sand shear angle > 32

and c

0, without any signicant need of extra compaction after hydraulic

placement.
The sand was transported from a temporary sand stock on land, by means
of a sand pump, to a sand spreader vessel. The vessel was provided with a fall
pipe with a 12-m wide horizontal spreader beam. With this system, adjusting
the sand ow and the dynamic positioning of the vessel, sand layers with more
or less uniform thicknesses could be placed.
After the installation of all Deep Mixing columns (already described in
section 8.6.2), a rst leveling sand layer with a thickness of about 1 m was
placed over the whole area of construction (see Fig. 9.1).
The rst embankment layer was built on top of the leveling layer. Also
the rst geotextile was placed and anchored on geocontainers as illustrated
in gure 9.1. The geocontainers (approximate dimensions 3 m2 m30 m),
lled with a mix of sand and and asphalt, were manufactured at the shoreline
and the carefully placed by means of a oating crane provided with state of
the art positioning systems (Fig. 9.2).
These operations were repeated for each 2 m sand layer to built up the
embankment. Today, the embankment has successfully reached the water level.
It is possible now, after about 2 years of construction, to see it emerging from
the water (Fig. 9.3).
9.3 Quality control of the embankment sand
Quality control of the embankment sand was performed regularly at several
stages during the construction by means of CPT tests. Out of cone penetration
tests it was possible to observe the state of the hydraulically placed sand with
depth. Moreover, some correlations of shear angle () and relative density
were attempted.
CPT tests were performed at several locations within the working area.
Figures 9.4 and 9.5 show some examples of CPT results on sand overlying
the SSI improved foundation soil. The gures show the quality control of the
Quality control of the embankment sand 125
Figure 9.2. Placing of geocontainers
Figure 9.3. Crown of the embankment beginning to emerge from the water
(January, 2006)
embankment sand up to a level close to TAW 0.00 which is more or less the
situation illustrated in gure 9.1.
Similarly, gures 3 and 4 show some examples of CPT results on sand
overlying the non-improved (soft) foundation soil in between SSI improved
soil areas.
126 Construction and monitoring of embankment
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
2
8
3
0
3
2
3
4
3
6
3
8
4
0
4
2
4
4

p
'

(

)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
2
0
4
0
6
0
8
0
1
0
0
D
r

(
%
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
2
4
6
8
1
0
q
c

(
M
P
a
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
Figure 9.4. CPT test on sand overlying the SSI improved foundation soil,
example 1
Overall, the cone penetration pressure q
c
is observed to increase reaching
values slightly greater than 10 MPa. However two dierent patterns can be
identied.
Sand overlying the improved zone shows low q
c
values at the interface with
the foundation layer (TAW -16.00) and then it increases to reach maximum
values at about 3 to 4 meters above such interface (close to TAW -13.00).
These patterns are probably caused due to the arching eect taking place
because of the presence of SSI treated columns. The arching eect causes the
Quality control of the embankment sand 127
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
2
8
3
0
3
2
3
4
3
6
3
8
4
0
4
2
4
4

p
'

(

)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

s
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
2
0
4
0
6
0
8
0
1
0
0
D
r

(
%
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

s
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
2
4
6
8
1
0
q
c

(
M
P
a
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

s
e
r
i
e
s
5
t
h

S
e
r
i
e
s
Figure 9.5. CPT test on sand overlying the SSI improved foundation soil,
example 2
sand to be most stressed some distance above the interface with foundation
layer while the sand below is barely receiving any surcharge.
On the other hand, sand overlying the non-improved foundation soil where
there are no SSI columns show a more regular pattern of q
c
with depth. In
fact, an almost linear trend was observed.
The sand shear angle correlated from CPT complies, in all cases, with the
design requirement of

=32

.
128 Construction and monitoring of embankment
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
2
8
3
0
3
2
3
4
3
6
3
8
4
0
4
2
4
4

p
'

(

)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
2
4
6
8
1
0
D
r

(
%
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
2
0
-
1
8
-
1
6
-
1
4
-
1
2
-
1
0
-
8
-
6
-
4
-
2 0
0
4
2
8
6
1
0
q
c

(
M
P
a
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
Figure 9.6. CPT test on sand overlying the non-improved foundation soil,
example 1
9.4 Instrumentation and monitoring
Figure 9.8 illustrates the layout of the installed instrumentation to contin-
uously follow up the excess pore water pressure and settlements during con-
struction.
Piezometers were installed at various locations within the SSI-improved
zones and the non-improved zones. They were also installed at dierent levels
within the foundation layer.
Instrumentation and monitoring 129
-
1
9
-
1
7
-
1
5
-
1
3
-
1
1
-
9
-
7
-
5
-
3
-
1 1
2
8
3
0
3
2
3
4
3
6
3
8
4
0
4
2
4
4

p
'

(

)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
9
-
1
7
-
1
5
-
1
3
-
1
1
-
9
-
7
-
5
-
3
-
1 1
0
2
0
4
0
6
0
8
0
1
0
0
D
r

(
%
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
-
1
9
-
1
7
-
1
5
-
1
3
-
1
1
-
9
-
7
-
5
-
3
-
1 1
0
2
4
6
8
1
0
1
2
q
c

(
M
P
a
)
TAW
1
s
t

S
e
r
i
e
s
2
n
d

S
e
r
i
e
s
3
r
d

S
e
r
i
e
s
4
t
h

S
e
r
i
e
s
5
t
h

S
e
r
i
e
s
Figure 9.7. CPT test on sand overlying the non-improved foundation soil,
example 2
Similarly, exible tubes were installed at 4 locations across the dock (at the
interface between embankment sand and foundation layer) to monitor vertical
displacements by means of water head dierences measured with a probe that
is displaced inside each tube. Out of this method it was possible to evaluate
settlement proles within the SSI-improved zone and the non-improved zone.
9.4.1 Excess pore water pressure
Figure 9.9 summarizes the measurements of excess pore water pressure (PWP)
in the foundation soil during construction. Signicant dierences can be
130 Construction and monitoring of embankment
Figure 9.8. Plan view of the construction site showing a layout of the installed
instrumentation
0
20
40
60
80
100
120
140
160
180
200
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
E
x
c
e
s
s

P
W
P

(
k
P
a
)
Underwater construction
works (Phase1)
Consolidation
Excess PWP in
the soft soil
deposit under
the embankment
Excess PWP in
the SSI improved
zone (between
SSI columns) at
the toes of the
embankment
Figure 9.9. Measurements of excess pore water pressure during construction
observed between measurements within the SSI-improved zone and the non-
improved zone.
As expected, the excess PWP in the improved zone is considerably smaller
than in the non-improved zone, showing that the SSI columns are indeed tak-
ing up an important portion of the load. On the other hand, the excess PWP
in the non-improved zone closely follows the stage construction loading history
and it is only after about a year of construction that signicant consolidation
slowly takes place.
Instrumentation and monitoring 131
S
e
t
t
l
e
m
e
n
t

w
i
t
h
i
n

t
h
e

n
o
n

i
m
p
r
o
v
e
d

f
o
u
n
d
a
t
i
o
n

z
o
n
e

(
p
r
o
f
i
l
e

Z
2
)
-
2
-
1
.
5
-
1
-
0
.
5 0
0
.
5 1
0
5
0
1
0
0
1
5
0
2
0
0
2
5
0
3
0
0
3
5
0
4
0
0
4
5
0
5
0
0
H
o
r
i
z
o
n
t
a
l

d
i
s
t
a
n
c
e

(
m
)
R e l a t i v e s e t t l e m e n t ( m )
3
5

d
a
y
s
1
2
0

d
a
y
s
2
2
0

d
a
y
s
2
9
0

d
a
y
s
3
9
0

d
a
y
s
4
4
0

d
a
y
s
5
0
0

d
a
y
s
5
7
0

d
a
y
s
6
4
0

d
a
y
s
7
6
0

d
a
y
s
N
W
S
E
D
r
e
d
g
e
d
m
a
t
e
r
i
a
l
S
e
t
t
l
e
m
e
n
t

t
u
b
e
L
t
u
b
e
=

1
0
0
m
L
t
u
b
e
=

4
5
0
m
F
i
g
u
r
e
9
.
1
0
.
S
e
t
t
l
e
m
e
n
t
p
r
o

l
e
a
t
t
h
e
i
n
t
e
r
f
a
c
e
b
e
t
w
e
e
n
e
m
b
a
n
k
m
e
n
t
s
a
n
d
a
n
d
f
o
u
n
d
a
t
i
o
n
l
a
y
e
r
w
i
t
h
i
n
t
h
e
n
o
n
-
i
m
p
r
o
v
e
d
z
o
n
e
132 Construction and monitoring of embankment
S
e
t
t
l
e
m
e
n
t

w
i
t
h
i
n

t
h
e

S
S
I
-
i
m
p
r
o
v
e
d

z
o
n
e

(
p
r
o
f
i
l
e

Z
4
)
-
2
-
1
.
5
-
1
-
0
.
5 0
0
.
5 1
0
5
0
1
0
0
1
5
0
2
0
0
2
5
0
3
0
0
3
5
0
4
0
0
4
5
0
5
0
0
H
o
r
i
z
o
n
t
a
l

d
i
s
t
a
n
c
e

(
m
)
R e l a t i v e s e t t l e m e n t ( m )
3
5

d
a
y
s
1
2
0

d
a
y
s
2
2
0

d
a
y
s
2
9
0

d
a
y
s
3
9
0

d
a
y
s
4
4
0

d
a
y
s
5
0
0

d
a
y
s
5
7
0

d
a
y
s
6
4
0

d
a
y
s
7
0
0

d
a
y
s
N
W
S
E
D
r
e
d
g
e
d

m
a
t
e
r
i
a
l
S
e
t
t
l
e
m
e
n
t

t
u
b
e
L
t
u
b
e

=

1
0
0
m
L
t
u
b
e

=

4
5
0
m
F
i
g
u
r
e
9
.
1
1
.
S
e
t
t
l
e
m
e
n
t
p
r
o

l
e
a
t
t
h
e
i
n
t
e
r
f
a
c
e
b
e
t
w
e
e
n
e
m
b
a
n
k
m
e
n
t
s
a
n
d
a
n
d
f
o
u
n
d
a
t
i
o
n
l
a
y
e
r
w
i
t
h
i
n
t
h
e
S
S
I
i
m
p
r
o
v
e
d
z
o
n
e
Instrumentation and monitoring 133
-3
-2.5
-2
-1.5
-1
-0.5
0
10 100 1000
Time (days)
S
e
t
t
l
e
m
e
n
t

(
m
)
Z2: Settlement of the soft
soil deposit under the
embankment
Z4 Settlement of the SSI-
improved zone under
the embankment
Z2
Z4
Estimated final settlement underthe current load
Figure 9.12. Settlement at the SSI-improved and nonimproved foundation zones
At this moment a more extended waiting period is being allowed so that
enough strength is gained (due to consolidation) in the foundation soil to
continue with the construction activities of the second phase.
9.4.2 Settlements
Figures 9.10 and 9.11 show the measured settlements proles in the non-
improved and the SSI improved foundation zones respectively. As expected,
the largest settlements are observed in the non-improved area were up to now
a maximum settlement of the order of 1.2 m was measured. On the other hand,
the maximum measured settlement in the SSI-improved zone is of the order
of 0.6 m.
Figure 9.12 compares the average settlements measured in both zones, SSI-
improved and nonimproved. It can be observed that the current settlement
in the nonimproved zone is about half the total settlement estimated for the
present surcharge.
134 Construction and monitoring of embankment
9.5 Conclusions
The construction of the sand embankment by hydraulic placement was shown
to produce a sand mass with enough shear strength and cone penetration
pressure to full the design requirements.
Excess PWP and settlement measurements so far show values in agree-
ment with the predictions at the design stage. In fact, the slow pore pressure
dissipation coupled to the fast settlement development (as observed in the
actual measurements) was already predicted with nite strain consolidation
theories. This shows that when taking account of Key aspects of soil behavior,
it is possible to properly model a complex problem such as this one and also
to obtain acceptable predictions.
References
G. AAS, Stability of natural slopes in quick clays. Proc. 10th Int. Conf. Soil
Mechanics and Foundation Engineering. Stockholm (1981).
J.H. ATKINSON, P.L. BRANSBY, The mechanics of soil: An introduction to crit-
ical state soil mechanics. McGraw-Hill, UK (1978).
L. BARDEN (1963) Stresses and displacements in a cross anisotropic soil.
Geotechnique, vol. 13, 198210.
L. BARDEN (1965) Consolidation of clay with non-linear viscosity. Geotechnique,
vol. 15, no. 4, 345362.
R.A. BARRON (1948) Consolidation of ne grained soils by drain wells. Transac-
tion, ASCE, vol. 113, 718742.
S. BERNANDER, Active earth pressure built-up: A trigger mechanism in large
landslides in sensitive clays. Technical report 49T, University of Lule (1981).
M.A. BIOT (1941) General theory of three dimensional consolidation. Journal of
Applied Physics, vol. 12, no. 2, 155164.
A.W. BISHOP, N. MORGENSTERN (1960) Stability coecients for earth slopes.
Geotechnique, vol. 10, 129150.
L. BJERRUM (1967) Engineering geology of Norwegian normally-consolidated
marine clays as related to settlements of buildings. Geotechnique, vol. 17, 81118.
L. BJERRUM, Eect of rate of strain on undrained shear strength of soft clays.
Proc. 7th Int. Conf. on Soil Mechanics and Foundation Engineering, Mexico
(1969).
L. BJERRUM, Problems of soil mechanics and construction on soft clays, SOA
report. Proc. 8th Int. Conf. on Soil Mechanics and Foundation Engineering,
Moscow, URRS, vol. 3 (1973), 111159.
A. BOUAZZA, W.F. VAN IMPE, W. HAEGEMAN (1996) Some mechanical
properties of reconstituted Boom Clay. Jour. Geot. & Geol. Engineering, vol. 14.
B. BROMS, Progressive translationary landslides. Swedish Geotechnical Institute.
Report No. 17 (1983).
D.A. BRUCE (2001) Practitioners guide to the deep mixing method. Ground
Improvement, vol. 5, no. 3, 95100.
K.W. CARGILL (1984) Prediction of consolidation of very soft soil. Journal of
Geotechnical Engineering, ASCE, vol. 110, no. 6, 775795.
COASTAL DEVELOPMENT INSTITUTE OF TECHNOLOGY, The deep mixing
method: principle, design and construction. A.A. Balkema, Rotterdam (2002).
E.H. DAVIES, G.P. RAYMOND (1965) A nonlinear theory of consolidation.
Geotechnique, vol. 15, 165173.
136 References
E. DE BEER, Shear strength characteristics of the Boom clay. Proc. of the
Geotechnical Conf. on Shear strength properties of natural soils and rocks,
Oslo, vol. 1 (1967).
P. DE WOLF, R. CARPENTIER, J. ALLAERT, J. DE ROUCK, Ground
improvement for the construction of the new outer harbour at Zeebrugge
Belgium. Proc. 8th Int. Conf. on Soil Mechanics and Foundation Engineering,
Helsinki, vol. 2 (1983), 827832.
J.M. DUNCAN, State of the art: Static stability and deformation analysis. ASCE
Geotechnical special publication No. 31, Stability and performance of slopes
and embankments II, vol. 1 (1992), 222266.
J.M. DUNCAN (1996) State of the art: Limit equilibrium and nite-element
analysis of slopes. Jour. Geotechnical Engineering Division ASCE, vol. 122, no.
7, 577596.
J.M. DUNCAN, S.G. WRIGHT, Soil strength and slope stability. John Wiley &
Sons, USA (2005).
R. DYVIK, C. MADSHUS, Laboratory measurements of G
max
using bender
elements. Proccedings of the Conference on advances in the art of testing soils
under cyclic conditions. Detroit (1985), 186196.
P.J. FOX, J.D. BERLES (1997) CS2: A piecewise-linear model for large strain
consolidation. Int. Jour. for Numerical Analysis in Geomechanics, vol. 21, no.
7, 453475.
A. GABERC, Increase of subsoil bearing capacity beneath embankments. Proc.
13th Int. Conf. on Soil Mechanics and Foundation Engineering, New Delhi,
India (1994), 759762.
R.E. GIBSON, G.L. ENGLAND, M.J.L. HUSSEY (1967) Theory of one-
dimensional consolidation of saturated clays. Geotechnique, vol. 17, no. 3,
261273.
R.E. GIBSON, R.L. SCHIFFMAN, K.W. CARGILL (1981) Theory of one-
dimensional consolidation of saturated clays, 2: Finite nonlinear consolidation
of thick homogeneous layers. Canadian Geotechnical Journal, vol. 18, no. 2,
s 280293.
O. GREGERSEN, The quick clay landslides in Rissa, Norway. NGI publication,
Oslo (1981).
W. HAEGEMAN, The measurement of small strain stiness of terciary sands and
clays with bender element tests in the triaxial apparatus. Proc. 13th Young
Geotechnical Engineers Conf., Santorini, Greece (1999).
S. HANSBO (1979) Consolidation of clay by band-shaped prefabricated drains.
Ground Engineering, vol. 12, no. 5, 1625.
D.J. HENKEL (1960) The shear strength of saturated remoulded clays. ASCE
Spec. Conf. on Shear Strength of Cohesive Soils, Boulder, 533554.
K. ISHIHARA (1993) Liquefaction and ow failure during earthquakes. Geotech-
nique, vol. 43, no. 3, 351415.
M. JAMIOLKOWSKI, C.C. LADD, J.T. GERMAINE, R. LANCELLOTTA, New
developments in eld and laboratory testing. Proc. 11th Int. Conf. on Soil
Mechanics and Foundation Engineering, San Francisco, vol. 1 (1985), 57153.
R.J. JARDINE, C.O. MENKITI, The undrained anisotropy of K0 consolidated
sediments. Geotechnical engineering for transportation infrastructure, Barends
et al. (eds.), Balkema, Rotterdam (1999).
D.C. KOUTSOFTAS, Undrained shear behaviour of a marine clay. Laboratory
shear strength of soil, ASTM STP 740, Yong & Towsend (eds.) (1981), 254276.
References 137
D.C. KOUTSOFTAS, C.C. LADD (1985) Design strengths for an oshore clay.
Jour. Geotechnical Engineering Division ASCE, vol. 111, no. 3, 337355.
R. LANCELLOTTA, Geotechnical engineering. AA Balkema, Rotterdam,
Netherlands (1995).
C.C. LADD (1991) 22nd Karl Terzaghi Lecture: Stability evaluation during staged
construction. Jour. Geotechnical Engineering Division ASCE, vol. 117, no. 4,
540615.
C.C. LADD, R. FOOT (1974) New design procedure for stability of soft clays.
Jour. Geotechnical Engineering Division ASCE, vol. 100, no. GT7, 763785.
Z. LECHOWICZ, An evaluation of the increase in shear strength of soft soils.
Advances in understanding and modelling the mechanical behaviour of peat,
den Haan, Termaat & Edil (eds.), Balkema, Rotterdam (1994), 167178.
S. LEROUEIL, J.P. MAGNAN, F. TAVENAS, Embankments on soft clays. Ellis
Horwood Limited, England (1990).
T. MATSUI, KA-CHING, Availability of shear strength reduction technique.
ASCE Geotechnical special publication No. 31, Stability and performance of
slopes and embankments II, vol. 1 (1992), 445460.
G. MESRI ABDEL-GHAFFAR (1993) Cohesion intercept in eective stress
stability analysis. Jour. Geotechnical Engineering Division ASCE, vol. 119, no.
8, 12291249.
J.K. MITCHELL, R.G. CAMPANELLA, A. SINGH (1968) Soil creep as a rate
process. Journal of the Soil Mechanics and Foundations Division, ASCE, vol.
94, no. SM1, 231253.
J. NUYENS et al., National Report 10 CPT in Belgium in 1995. Proc. Int.
Symposium on Cone Penetration Test, Sweden, vol. 1 (1995), 1727.
I. ODLER, Special inorganic cements. E&FN Spon, London (2000).
A. PORBAHA, S. SHIBUYA, T. KISHIDA (2000) State of the art in deep mixing
technology. Part III: Geomaterial characterization. Ground Improvement, vol.
4, no. 3, 91110.
I.C. PYRAH (1996) One-dimensional consolidation of layered soils. Geotechnique,
vol. 46, no. 3, 555560.
H. RATHMAYER (1996) Deep mixing method for soft subsoil improvement in the
Nordic countries, Grouting and Deep Mixing. Proc. 2nd Int. Conf. on Ground
Improvement Geosystems, Balkema, 2, 869877.
P.K. ROBERTSON, C.E. WRIDE (1998) Evaluating cyclic liquefaction potential
using cone penetration test. Canadian Geotechnical Journal, vol. 35, no. 3,
442459.
P.W. ROSCOE, A.N. SCHOFIELD, C.P. WROTH (1958) On the yielding of soils.
Geotechnique, vol. 1, 2252.
S. SAITOH, Y. SUZUKI, K. SHIRAI (1985) Hardening of soil improved by deep
mixing method. Proc. Int. Conf. on Soil Mechanics and Foundation Engineering,
11, San Francisco, Aug. 1985. vol. 3.
R.L. SCHIFFMAN (1980) Finite and innitesimal strain consolidation. Journal of
the Geotechnical Engineering Division, ASCE, vol. 106, no. GT2, s 203207.
R.L. SCHIFFMAN, Theories of consolidation: a comparative study. Proc. Sym-
posium on Developments in Geotechnical Engineering, Bangkok, Thailand,
January (1994).
R.L. SCHIFFMAN, S.K. ARYA, One-dimensional consolidation. Numerical
methods in geotechnical engineering, Desai & Christian (eds.), McGraw-Hill
(1977), 364398.
138 References
A. SCHOFIELD, P. WROTH, Critical state soil mechanics. McGraw-Hill, UK
(1968).
H.B. SEED, I.M. IDRISS (1971) Simplied procedure for evaluating soil liquefac-
tion potential. Journal of the Soil Mechanics and Foundations Division, ASCE,
vol. 97, no. SM9, 12491273.
F. TAVENAS, Some aspects of clay behaviour and their consequences on modelling
techniques. Laboratory shear strength of soil, ASTM STP 740; Yong, Towsend
(eds.) (1981), 667677.
F. TAVENAS, S. LEROUEIL (1987) State of the art on laboratory and in-situ
stress-strain-time behaviour of soft clays. Proceedings of International sympo-
sium on geotechnical engineering of soft soils, Ciudad de Mexico, Agosto 1987,
vol. 2.
D.W. TAYLOR (1937) Stability of earth slopes. Journal of Boston Society of Civil
Engineering, vol. 24, 197246.
K. TERZAGHI, Erdbaumechaniek auf bodenphysikaliser Grundlage. Leipzig
Deuticke (1925).
USACE (1990) Engineering and Design Settlement Analysis. Publication
Number: EM 1110-1-1904.
P. VAN IMPE, Consolidatie van verzadigde, sterk samendrukbare poreuze media.
Final degree work, Faculty of applied sciences, University of Ghent (1999).
W.F. VAN IMPE, Foundation engineering problems at the construction of the new
Zebrugge harbour. Internal report, Laboratory of soil mechanics, University of
Ghent (1985).
W.F. VAN IMPE, Evaluation of deformation and bearing capacity parameters of
foundations from static CPT results. Proc. 4th Int. Geot. Seminar, Singapore
(1986).
W.F. VAN IMPE, Soil improvement techniques and their evolution. AA Balkema,
Rotterdam, Netherlands (1989).
W.F. VAN IMPE (1997) Soil improvement experiences in Belgium: part III. Case
histories. Ground Improvement, vol. 1, no. 4, 179191.
P.O. VAN IMPE (1999) Consolidation of very compressible saturated porous media.
Final degree work, Faculty of Applied Sciences, Ghent University. In Dutch.
W.F. VAN IMPE, Deep mixing. Internal report, Laboratory of soil mechanics,
University of Ghent (2000).
W.F. VAN IMPE, E. DE BEER (1984) Contribution to the analysis of landslides in
quick clay. International symposium on landslides, 4, Toronto, Sept. 1984. Vol. 2.
W.F. VAN IMPE, R.D. VER

ASTEGUI FLORES, M. VAN DEN BROECK, P.

MENG

E (2004) Stabilisation of dredged sludge with dierent types of binders.

Proc. Int. Symp. on Engineering Practice and Performance of Soft Deposits,
Osaka, 197200.
R.D. VER

ASTEGUI FLORES (2001) Geotechnics of staged construction. Final

degree work, Faculty of Applied Sciences, Ghent University.
R.D. VERASTEGUI FLORES et al., Lime-cement mixing for improvement
of alluvial soft Belgian soil. Proc. Int. Symp. on Engineering Practice and
Performance of Soft Deposits, IS-OSAKA (2004), Japan 241246.
R.D. VERASTEGUI FLORES, W.F. VAN IMPE, P. AFSCHRIFT, W.
CROMHEEKE (2005) Lime cement columns in alluvial soft soil. Proc.
Int. Conf. Soil Mechanics and Geotechnical Engineering, 16, Osaka, vol. 3.
Subject Index
behavior of clay 35, 1116
bender element test 11213
cone penetration test 50, 526, 72, 98,
106, 124, 1269
consolidation analysis 7, 29, 62, 64
constant rate of strain consolidation
test 778
critical state 1415, 79
deep mixing 2, 77, 81, 91122
dilatometer test 29, 989, 1067
nite strain theory 57, 647, 6970, 77,
84, 134
geotextile reinforcement 81, 83, 86, 88,
1234
ground improvement 2, 91122
hydration of cement 945
hydration of lime 94
innitesimal strain theory 8, 57, 5864,
66, 6970, 84
k
0
-consolidated clay 12, 15, 203
limit equilibrium method 25, 30, 83
liquefaction assessment 4950
normalized behaviour 1314, 18, 76
normally consolidated clay 46, 11, 18
on-land deep mixing 97109
overconsolidated clay 4, 1415, 72,
78, 83
scanning electron microscopy 734, 75,
119, 1212
sensitive clays 27, 3549
slope failure patterns, 256
soft soil improvement (SSI) 2, 523,
110, 11719, 125133
soft soils 48
staged construction 5, 8, 20, 28, 67,
70, 85
stiness of treated soil 107, 11214
strength reduction method 25, 313, 35,
51, 83
stress induced anisotropy 20
structural anisotropy 20
temperature eect on hydration 11416
unconned compressive strength 1001,
104, 111, 11516, 119
underwater deep mixing 10922
underwater embankment 4, 67, 819,
1235
undrained shear strength 7, 1124, 29,
76, 83, 88, 99
undrained strength anisotropy 202
undrained strength ratio 1314, 19, 22,
76