Numerical Analysis of Cantilever and Anchored Sheet Pile Walls

Terms and Conditions of Use:
this document downloaded from
vulcanhammer.net
Since 1997, your complete
online resource for
information geotecnical
engineering and deep
foundations:
The Wave Equation Page for
Piling
Online books on all aspects of
soil mechanics, foundations and
marine construction
Free general engineering and
geotechnical software
All of the information, data and computer software

(information) presented on this web site is for general
information only. While every effort will be made to insure
its accuracy, this information should not be used or relied on
for any specific application without independent, competent
professional examination and verification of its accuracy,
suitability and applicability by a licensed professional. Anyone
making use of this information does so at his or her own risk
and assumes any and all liability resulting from such use.
The entire risk as to quality or usability of the information
contained within is with the reader. In no event will this web
page or webmaster be held liable, nor does this web page
or its webmaster provide insurance against liability, for
any damages including lost profits, lost savings or any
other incidental or consequential damages arising from
the use or inability to use the information contained
within.
This site is not an official site of Prentice-Hall,
Pile Buck, the University of Tennessee at
Chattanooga, or Vulcan Foundation
Equipment. All references to sources of
software, equipment, parts, service
or repairs do not constitute an
endorsement.
And much more...
Visit our
companion site
http://www.vulcanhammer.org
MASTERTHESIS
Master
CIVIL ENGINEERING
Title
NUMERICAL ANALYSIS OF CANTILEVER AND

ANCHORED SHEET PILE WALLS AT FAILURE AND
COMPARISON WITH CLASSICAL METHODS
Author
ALEJO GONZLEZ TORRABADELLA
Tutors
ALEJANDRO JOSA GARCA-TORNEL

SEBASTI OLIVELLA PASTALLE
Speciality
GEOTECHNICS
Date
JANUARY 2013

Acknowledgments. Agradecimientos
Gracias a mi familia, a mis compaeros y a mis amigos. Gracias a mi madre por

apoyarme, no slo en esta tesina, sino en todo lo que he hecho. Gracias a Guille por
compartir tantas horas de biblioteca. Gracias a Jose, por tu ayuda y tambin por las
horas compartidas. Gracias a todos los compaeros de mster por la buena sintona.
Gracias a Agustn: esta tesina es la continuacin de la tuya y he pretendido estar a la
altura de un listn que dejaste alto. Gracias por estar siempre disponible y por tu
inestimable ayuda.
Por ltimo gracias a mis tutores, Alejandro y Sebasti, no slo por su papel crucial en
la direccin de este estudio, tambin por las lecciones aprendidas en los aos de Obras
Pblicas.
Numerical analysis of cantilever and anchored sheet pile walls at failure and comparison with classical methods
INDEX
1. INTRODUCTION AND OBJECTIVES .................................................................. 3

1.1.
Introduction ........................................................................................................ 3
1.2.
Objectives .......................................................................................................... 3
1.3.
Contents ............................................................................................................. 4
2. BACKGROUND....................................................................................................... 5
2.1.
First sight on sheet pile walls ............................................................................. 5
2.2.
Global mechanical behaviour ............................................................................ 6
2.2.1.
2.3.
Stress state .................................................................................................. 6
Earth pressures ................................................................................................... 8
2.3.1.
Coulomb Theory ......................................................................................... 8
2.3.2.
Rankine Theory .......................................................................................... 9
2.4.
Classical design methods ................................................................................. 10
2.4.1.
Cantilever walls ........................................................................................ 11
2.4.2.
Single-anchored walls .............................................................................. 13
2.4.3.
Multiple-anchored walls ........................................................................... 15
3. BIBLIOGRAPHIC REVIEW ................................................................................. 18

3.1.
Free embedded cantilever sheet pile walls....................................................... 18
3.2.
Anchored sheet pile walls ................................................................................ 29
4. NUMERICAL MODEL .......................................................................................... 34

4.1.
Introduction ...................................................................................................... 34
4.2.
Modelling sheet pile walls ............................................................................... 35
4.2.1.
Geometry .................................................................................................. 36
4.2.2.
Loads and boundary conditions ................................................................ 38
4.2.3.
Material properties.................................................................................... 38
4.2.4.
Mesh and initial stresses generation ......................................................... 41
5. INFLUENCE OF THE INITIAL STRESS STATE ............................................... 44

5.1.
Methodology .................................................................................................... 45
5.2.
Analytical methods .......................................................................................... 46
5.3.
Results analysis ................................................................................................ 46
6. MULTIPLE-ANCHORED SHEET PILE WALLS ................................................ 49

6.1.
Introduction ...................................................................................................... 49
1
6.2.
Geometry and methodology ............................................................................. 51
6.2.1.
Study of the influence of the construction procedure ............................... 51
6.2.2.
Study of anchor load variation ................................................................. 52
6.2.3.
Three-level-anchored sheet pile wall ........................................................ 53
6.3.
Results analysis ................................................................................................ 54
7. CONCLUSIONS ..................................................................................................... 62
8. REFERENCES ........................................................................................................ 64
ANNEX I: PAPER ......................................................................................................... 68
ANEEX II: BIBLIOGRAPHIC REVIEW ..................................................................... 91
ANNEX III: STUDY OF THE K0 INFLUENCE .......................................................... 94
ANNEX IV: MULTIPLE-ANCHORED SHEET PILE WALLS ................................ 103
1. INTRODUCTION AND OBJECTIVES

1.1.
Introduction
Concrete sheet pile walls were seldom designed before de beginning of the 20th
Century, dating the first design methods from the early 1900s. It was in the 1950s,
when sheet pile walls were broadly established as a solution to solve problems
associated with deep excavations near buildings, subterranean structures or below the
water table. Since then, the growing need to use scarce land efficiently, along with the
improvement and development of specialized machinery with a greater efficiency, have
led to an increase in the use of sheet pile walls.
Although design methods have been constantly reviewed and improved, these have not
changed much in the last 50 years. Its usage is fully extended due to its simplicity and
reliability. Despite the development of numerical methods in the last decades applied to
geotechnical engineering the classical analytical methods are still broadly used.
This Master Thesis is framed into a wider study of the behaviour of sheet pile walls at
failure. In particular, it is the continuation of a previous work by Cuadrado (2010)
[Stress-strain analysis at failure and safety conditions in cantilever and anchored sheet
pile walls. Comparison with classical methods]. In this study the author developed a
detailed assessment of the classical methods for cantilever and single-anchored sheet
pile walls and compared them with the Finite Element method. Additionally, that work
included a contribution on safety practices, consisting of increasing the embedment
depth by 20% and reducing the passive resistance.
The first motivation of this Master Thesis is the elaboration of a scientific paper based
on the previous work of Cuadrado (2010). To do so, several weak points of this study
had to be reviewed. As a first step, the sate of knowledge had to be reviewed in depth.
The contribution of this master thesis has been to study the influence of some
parameters whose influence was previously missed. The analysis has been focused on
observing changes in the behaviour of sheet pile walls when those parameters or most
significant data are changed. The main findings are that the initial stress state of the soil
has an influence into the cantilever sheet pile wall behaviour. For K0 (at-rest coefficient)
values between 0.7 and 0.9 the minimum wall movements are registered.
The second motivation of this Master Thesis is the extension of the sheet pile wall
modelization to the multiple-anchored case. For that case, the effects of increasing the
anchor force and the staged excavation are analysed. Results show that when increasing
the anchor force the movement of the wall is limited. In the staged excavation study, it
may be seen that larger movements are registered with a late installation of the anchors.
The effect of the early pre-stressing of the anchor is then to reduce the movements,
particularly at the top of the wall.
1.2.
Objectives
As mentioned, the main objectives of this Master Thesis are the elaboration of a paper
towards the publication in a specialized journal, and to extend the previous studies to
multiple-anchored sheet pile walls. Additionally, and as previous steps, this work
thoroughly reviews the existing literature and validates the previous results.
3
The process has consisted of the following chronological steps:

- In-depth bibliographic review of published papers and all the available literature
related to sheet pile walls. This review has included design limit equilibrium
methods, numerical methods and theoretical or applied studies that have been widely
treated since the middle 20th century.
- Review and appraisal of the previous results focused on cantilever and singleanchored sheet pile walls. Cases have been replicated and the influence of previously
determined parameters has been studied in depth.
- Elaboration of a scientific paper, in order to be published according to a standardized
layout.
- Extension of the previous study to multiple-anchored sheet pile walls. Resolution of
the cases proposed through the Finite Element program Plaxis and analysis of the
results.
1.3.
Contents
This document can be seen as a collection of several fairly independent studies, which
however constitute the parts of a total single body.
This master thesis is organized as follows:
The second chapter is named Background and includes a general description of sheet
pile walls. It introduces the reader into the background of the sheet pile wall design,
starting from the grounds, the earth pressure calculation and the classical methods.
The third chapter contains a review of the state of knowledge, from the first publications
related to sheet pile walls until the most recent papers published in geotechnical
journals. Some publications are merely mentioned, some others are reviewed in depth.
The fourth chapter describes the general use of the commercial software Plaxis and
presents the model used.
The fifth chapter is born from the previous work of Cuadrado (2010), and discusses the
influence of the initial stress state on sheet pile walls behaviour.
The sixth chapter presents the study of a two-level-anchored sheet pile wall. A
discussion of the construction procedure and the anchor load is presented. This chapter
includes the modelization of a three-level-anchored sheet pile wall, and results are
compared with the two-level-anchored one.
Finally, the seventh chapter presents the conclusions of the study.
2. BACKGROUND
2.1.
First sight on sheet pile walls
Retaining structures are commonly divided into two families:

- Rigid retaining structures, where the stability is provided by the use of a large volume
of mass. Only rigid body movements can occur.
- Flexible retaining structures, where other properties of the materials, such as stiffness,
strength and wall thickness are exploited in order to provide the stability. Bending and
rigid body movements are found.
Sheet pile walls are classified as flexible retaining structures. The stability is provided
through an embedment of the wall on the ground working as a cantilever structure (see
Figure 2.1) and eventually a system of anchors (see Figure 2.2), so the wall is subject to
shear stresses and bending moments. One of the main benefits is the minimization of
used material, in contrast to the needs of rigid retaining structures.
Sheet pile walls are in essence walls built before excavating, with the double aim of
resisting the earth pressure generated by the excavation, and limiting or preventing the
water inlet. The execution of a sheet pile wall consists on an excavation of a trench,
large and deep, without the need of shoring, but usually stabilized with bentonite
slurries. Then, reinforcement bars are placed and the concrete is laid, or prefabricated
elements are placed instead.
Sheet pile walls are not only able to resist earth pressures and prevent the water inlet,
but also to receive vertical loads transmitted by other structural elements. Moreover,
sheet pile walls are an efficient solution to limit ground movements associated with the
excavation. Thus, there are four main functions that can be carried out by a sheet pile
wall (Jimnez Salas, 1980):
i)
ii)
iii)
iv)
Resist the thrust generated by the excavation.

Limit the movements in the back of the wall, in the unexcavated zone.
Prevent from the water inlet.
Support vertical loads.
Figure 2.1 Cantilever sheet pile wall
Figure 2.2 Anchored sheet pile wall. At first sight, lesser embedment is needed
2.2.
Global mechanical behaviour
The main feature of an embedded retaining structure is the contribution on stability of

the embedded zone. Hence, the main unknown parameter to define is the embedment
depth, which depends on the magnitude and distribution of the earth pressures over the
wall. Earth pressure acting on the sheet pile wall depends, at the same time, on:
-
The type of earth pressure distribution

Geometry
Flexibility of the sheet pile wall
Soil-structure interaction
Construction method
2.2.1. Stress state

The stress state in each point of the sheet pile wall depends on the movement produced.
In order to define the earth pressure at failure conditions, the Rankine states are used.
Supposing a horizontal soil surface, with no external loads, the effective horizontal and
vertical stress increase linearly with depth, and its relation is known as at-rest
coefficient (K0):
(2.1)
When a sheet pile wall is introduced and an excavation is done, the earth pressure
generated can be produced in two ways:
-
Active pressure (Ka): The soil exerts a pressure against the wall. The wall moves to
the excavation and the horizontal stress decreases, as the vertical stress remains
unchanged. A decompression in the horizontal stress occurs. In a limit situation, a
failure wedge is formed, producing a plastic regime. The earth pressure is lower
than in the at-rest state.
Passive pressure (Kp): The wall exerts a pressure against the soil. In this case the
horizontal stress increases, while the vertical stress remains unchanged. Therefore,
the earth pressure is higher than in the at-rest state. In a limit situation, a failure
wedge is formed as well, but with greater dimensions than in the active case.
Figure 2.3 shows the horizontal thrust acting on the sheet pile wall. The ordinate axis
can be replaced by K0 by dividing by v, according to eq. 2.1, the bigger the movement
6
is, the higher the horizontal stress is, and so the earth pressure depends essentially on
the strain. The asymptotes represent the active and the passive limit states, known as
Rankine limit states. Normally, K0 is closer to Ka than Kp, always meeting the following
condition:
Ka<K0<Kp
(2.2)
'h
passivestate
atreststate
activestate
movement
Figure 2.3 Basic scheme of horizontal stress depending on the movement. The vertical axis
may be replaced by K0 by dividing by v
The three-dimensional stress state in one point may be represented by Mohrs circle, a
two-dimensional graphical representation (see Figure 2.4). The stress state in one point
may be broken in two components, the normal stress (ordinate n), and the shear stress
(abscissa ). Principal stresses are I (major) and III (minor). The third principal stress
II, not represented, is assumed to be equal to III. The circumference of the circle is the
range of points that represent the state of stress on individual planes at all their
orientations. The major and minor normal stresses are given when the shear stress is
zero. The failure occurs when the linear envelope of Mohr Coulomb is surpassed (eq.
2.3). As seen in eq. 2.3 and Figure 2.4, the effect of cohesion is always favourable.
(2.3)
Figure 2.4 Mohrs circle. In this case, the failure criterion of Mohr Coulomb has been
reached.
2.3.
Earth pressures
In addition to the stress state in each point, estimating the earth pressure distribution is
critical in order to design a retaining structure. Regarding this point, different theories
are available. The theory applied to rigid structures can be applied, generally, to sheet
pile walls as well. The back of the wall is, in every case, vertical. However, in sheet pile
walls the passive pressure is more significant. Theories of Coulomb and Rankine are
reviewed below. Both are commonly accepted for calculating earth pressures. The first
dates from 1776 and the second from 1857. The general equations developed for both
theories are based on the fundamental assumptions that the retained soil is cohesionless,
homogeneous and drained.
2.3.1. Coulomb Theory

The main hypothesis of Coulomb is the creation of a flat failure surface of the soil that
produces a slippage plane. This assumption is appropriate only in the active case,
whereas in the passive case a significant error is found. Other hypotheses are the
homogeneity of the soil, with a soil friction angle and the fulfilment of the MohrCoulomb failure criterion in the slippage plane. In addition, the method requires a
sufficient rotation or yielding of the wall in order to engage the entire shear strength.
The method is essentially based on the equilibrium of wedges at failure. In the active
case, a maximum wedge to engage the maximum thrust is formed (see Figure 2.5),
whereas in the passive case it is a wedge to provide the minimum thrust. Under
conditions of homogeneous, cohesionless and drained soil, a linear earth pressure law is
considered corresponding to the active case (eq. 2.4).
(2.4)
Where Ka is:
(2.5)
Where the angles , and , as seen in Figure 2.5, represent:
: Inclination of the wall from the horizontal.

: Inclination of the ground in the back of the wall.
: Friction angle between the wall and the soil.
Figure 2.5 Coulomb wedge analysis
Notice that the interaction ground-structure promotes stability. The thrust magnitude is
calculated through the equilibrium of forces involved at failure, and its direction
depends directly on the angle adopted. The application point is assumed to be the
centre of mass of the multiple unitary thrusts in the back of the wall.
The Coulomb theory is appropriate for the active case, and it is known to be slightly
conservative. On the other hand, it is unsuitable for the passive case, especially for high
values of and , and that makes it critical for sheet pile walls. Its greatest benefit is the
versatility for complex geometries.
2.3.2. Rankine Theory

The main hypothesis of this theory is the condition of the soil to be in a Rankine limit
state. A Rankine limit state is a stress state of plastic equilibrium, where failure surfaces
are found in only two directions. In these surfaces the soil has reached the MohrCoulomb failure criterion (eq. 2.3).
As said, the failure surfaces in two directions define the size of the wedge produced. In
the passive case, the failure involves much more volume of mass than in the active case,
as seen in Figure 2.6.
a
B
Figure 2.6 Failure wedges. a) Passive limit state. b) Active limit sate.
Under this criterion, the failure occurs with only two soil friction angles, for
respectively active and passive limit state, as shown in Figure 2.7. These angles
determine the failure wedge formed, that will be different, as said, for the active and the
passive limit state. Notice that in the active state, the minor stress III corresponds to the
9
horizontal stress h, and the major stress I corresponds to the vertical stress v.
Accordingly, in the passive state, I=h and III=v. The failure surface in Figure 2.7 is
found by drawing a line from the pole to the point where the failure envelope is reached.
Figure 2.7 Mohr-Coulomb failure criterion. The black circle represents an initial stress
state where v>h. A cohesion c has been considered
The failure state can be reached therefore from either active or passive state. The
effective stress for each point of the soil, considering a horizontal ground surface, is:
(Active case)
(2.6)
(Passive case)
(2.7)
After a geometrical procedure, it can be proved that the angle of the failure surface, in
Figure 2.7, depends only on the soil friction angle (), being
and
for the active and passive limit state respectively. Coefficients Ka and Kp depend as well
on the soil friction angle (). Under conditions of homogeneous soil and horizontal
surface, its values are:
(2.8)
(2.9)
The Rankine theory is suitable for the active state, and it is known to be slightly
conservative. It is, nonetheless, inadequate for the passive state. Its greatest benefit is
the simplification achieved, especially for simple cases.
2.4.
Classical design methods
There are several design methods for sheet pile walls. The original proposals date back
from the first half of the 20th century, and have been constantly reviewed. Some of them
are still reviewed and studied nowadays.
Analytical methods can be found under the name classical methods or limit-stage
design methods (King, 1995). Regardless the name adopted, geotechnical design
calculations in either cantilever or anchored sheet pile walls establish equilibrium of
10
horizontal forces and moments to define the failure state and the reference embedment
below the dredge line (before applying the safety factor). The main analytical methods
are presented below.
2.4.1. Cantilever walls

In cantilever walls, a reaction force appears at the bottom of the wall, allowing stability,
as shown in Figure 2.8. When the embedment depth is increased, further reaction
appears in the opposite direction, but this is strictly unnecessary to guarantee stability.
Figure 2.8 Schematic vision of the net earth pressure in a cantilever wall
Stability in cantilever sheet pile walls depends on an adequate embedment below the
dredge line. The classical limit equilibrium methods attempt to model the sheet pile
walls at failure conditions, and differ from each other in several assumptions, but being
a common feature the reach of the failure state in the whole length of the wall.
The main analytical methods for cantilever walls (full method, simplified method and
gradual method) are reviewed below.
Full method
This method, shown in Figure 2.9 and known as UK method as well, has been fully
described by Padfield & Mair (1984) and gets its name in contrast to the simplified
method, described later. The active limit state is assumed to be reached in the back of
the wall above the rotation point, and the passive limit state is assumed to be reached in
front of the wall between the dredge line and the rotation point. An overturn in the
normal pressure direction is supposed to be produced at the rotation point, below which
the full passive pressure is moved behind the wall and the active to the front, so there is
a sudden jump in the earth pressure distribution which is needed to prescribe moment
equilibrium.
11
Figure 2.9 Full method. Note the sudden change in the earth pressure distribution
Simplified method
Due to the complexity of the full method, a simplification is recommended (Padfield &
Mair, 1984). As shown in Figure 2.10, the earth pressure below the rotation point can be
replaced by an equivalent concentrated force acting on point O, represented as R.
The value of d calculated is considerably lower than the one calculated from the full
method. The common practice is to increase it by 20%, being therefore the reference
embedment length equal to 1,2d. The zone below the rotation point is assumed to
receive the passive earth pressure, simplified as R, and a verification must be done in
order to ensure that the additional length 0,2d is able to generate the value of R.
The simplified method is slightly more conservative than the other methods, although it
leads to appreciably similar results (Padfield & Mair, 1984). Its greatest benefit is the
0,
0).
simplicity achieved on the traditional system of equations (
Figure 2.10 Simplified method. The thrust below the rotation

point O is replaced by a concentrated force R
Gradual method
A slightly different approach was made by Krey (1932), and later reviewed by Bowles
(1988), which does not involve the hypothesis of a sudden change in the earth pressure
distribution. The assumption made in this method is to consider a transition zone where
the net earth pressure gradually changes its direction from the front to the back of the
wall, as shown in Figure 2.11. This transition occurs around the rotation point and is
assumed to be linear. The earth pressure at the bottom of the wall is known (see Fig.4),
12
so the incognita to obtain the embedment depth is the height z. The gradual method is
known as USA method (Day, 1999; krabl, 2006) or general rectilinear net pressure
method as well (Day, 2001). All of them consist of modern versions of the method
initially proposed by Krey in 1932.
Figure 2.11 Gradual method, shown in net pressures
2.4.2. Single-anchored walls

Anchored walls (or tied-back walls) with a single row of anchors are able to achieve the
equilibrium without considering a passive reaction at the bottom of the back of the wall,
as seen in Figure 2.12. However, this reaction may be considered depending on the
design method. The tie rod position, its inclination and its load are the subject of
continuous studies.
The main advantage of an anchored sheet pile wall, against those cantilevered, is the
ability to reduce the embedment depth. In other words, it has the ability to increase the
excavation depth, which makes the structure more profitable. Some disadvantages are
found as well, related to higher amount of unknowns in the problem. Although the
excavation depth may be increased thanks to the existence of the anchor, it should not
be forgotten, that until the anchor is placed, the structure behaves as a cantilever sheet
pile wall.
Figure 2.12 Schematic vision of the net earth pressure in an anchored wall
13
The main analytical methods for anchored walls are presented below, the free earth
support method and the fixed earth support method, which differ on the hypotheses
adopted.
Free earth support method
This method is based on the assumption that movements on the embedded zone of the
wall are sufficient to mobilize the active and passive thrust behind and in front of the
wall respectively. The passive pressure is assumed to act only in front of the wall
through the depth d (Figure 2.13). The bottom of the wall has therefore free movement,
and a minimum reference embedment depth, to satisfy equilibrium, is obtained.
The equilibrium is fulfilled between the passive and active pressures, and the anchor
force, for obtaining the embedment depth. The way to proceed is to take moments
respect the point of application of the anchor (the anchor head) and then equating this
expression to zero. This equation provides the minimum embedment depth d to provide
equilibrium.
Figure 2.13 Free earth support method. The equilibrium involves movement of the bottom
Fixed earth support method

By increasing the embedment depth, an overturn in the normal earth pressure is
achieved, until the deflected shape at the bottom becomes approximately vertical.
Hence, the earth pressure distribution is essentially equal to the cantilever case (Figure
2.14), and the wall behaves like a partially built-in beam subjected to bending moments
(United States Steel, 1984).
14
Figure 2.14 Fixed earth support method
This method leads to a maximum reference embedment depth. It requires an additional

equation which is usually defined estimating where the zero net pressure (inflection
point O; zero bending moment) is located. It is frequent to locate such point in the
middle of the embedment depth, though other assumptions have been proposed. Blum
(1931) showed a dependence relation between the distance from the inflection point to
the dredge level and Ka.
The procedure to find the embedment depth is known as Blums equivalent beam
(shown in Figure 2.15), which consists of treating the sheet pile wall as a beam, and
dividing it in an upper beam, where net earth pressures act against the back of the wall;
and a lower beam, where net earth pressures act against the front of the wall. In the
upper beam, equilibrium of moments may be taken at the anchor point to find Rb. Then,
in the lower beam, moments are taken at the bottom to find the embedment depth,
which must be increased in order to ensure that a reaction Rc is able to be engaged.
Figure 2.15 Blums equivalent beam for the design of anchored sheet pile walls
2.4.3. Multiple-anchored walls

Multiple-anchored sheet pile walls represent a totally different geotechnical subject. The
complexity of the problem is increased as the number of anchors grows. In fact, no
analytical methods exist for the design of multi-anchored sheet pile walls. However, and
despite the inexistence of analytical methods, the use of this kind of structures has been
widely used. Owing to this, some analytical data are available, which contributes to a
better understanding of the behaviour of multiple-anchored sheet pile walls.
15
By anchoring the free height of the wall in multiple rows, the embedment depth may be
reduced significantly, which represents the most important advantage of this structure.
Being strict, no embedment depth is needed to guarantee the stability. The action of the
anchors is sufficient to provide equilibrium. However, by decreasing the embedment
depth may appear other problems, as global instability or instability of the bottom of the
excavation.
The design process consists of calculating the sheet pile wall as whether it was a
continuous beam. The earth pressure adopted is corresponding to the stress distribution
under the hypothesis of free earth support. All the anchors must receive tensile stress,
and never of compression. Whether compression occurs, a new location of the anchors
must be adopted. Frequently, trials must be done in an iterative process in order to
optimize the distribution of moments and anchor loads.
Even in the absence of analytical methods, an implicit hypothesis must be considered:
the absence of passive pressure in the zone of anchors. Otherwise, the structure must be
treated as a propped wall. As in single-anchored sheet pile walls, it should not be
forgotten that stability must be guaranteed in all the temporary phases during the
construction, and not only in the final situation.
There are some specific calculation procedures, as that proposed by Caquot, based on
locating the anchors in order to receive the same load, and exploiting the bending
strength of the sheet pile wall. This solution creates a non uniform distribution of the
anchors, where the spacing increases with the height. Other solutions involve the
arbitrary location of the anchors or its inclination.
Few literatures have been published in relation to multiple-anchored sheet pile walls.
Plant (1972) showed a series of tests in a multiple-anchored sheet pile wall. The effect
of the anchor inclination on wall movement, sand subsidence, anchor load changes and
earth pressure distribution were studied. A reduced model of sheet pile wall of 1 m high
with 3 rows of pre-stressed anchors was used (shown in Figure 2.16). The material used
was dry sand. The main conclusions from this study were (Plant, 1972):
- The behaviour of a multi-anchored retaining wall is complex and is related to the
interaction between the wall, the soil and the anchors.
- The effect of increasing the anchors inclination is to magnify the wall and soil
movements.
- The equivalent earth pressure coefficient1 at full excavation became smaller as the
anchor inclination increased.
- Consideration should be given to having the middle and lower levels of anchors
less steeply inclined than the upper level to provide a more rigid integrated
structural system in order to reduce movements.
Plant (1972) noticed the need to develop analytical methods, referring to finite
elements, applied to sheet pile walls for the first time.
The equivalent earth pressure coefficient
is an average relation between the horizontal
pressure acting against the back of the wall and the horizontal theoretical thrust. In this ratio, the
numerator is variable with the anchor inclination and the denominator remains unchanged.
16
Figure 2.16 Model of multi-anchored sheet pile wall carried out by Plant
17
3. BIBLIOGRAPHIC REVIEW
Literature referred to sheet pile walls is diverse and may confuse the reader. First
publications date from the early 20th century. Several authors have studied both
anchored and cantilever sheet pile walls, in the same or in different publications.
However, a distinction has been made in this study by separating the literature referred
to cantilever, anchored and multi-anchored sheet pile walls. After a deep survey, the
most relevant literature is summarized in Annex II. The main sources have been
journals such as Gotechnique, Proceedings of Institution of Civil Engineering,
Canadian Geotechnical Journal, Computers and Geotechnics; and on second term other
less known journals, some essential books and proceedings from international
workshops.
3.1.
Free embedded cantilever sheet pile walls
Engels method
Cantilever sheet pile walls were seldom designed before the beginning of the 20th
century. One of the first design methods was introduced by Engels (1903), based on
experimental observations. A parabolic net earth pressure distribution over the whole
depth of embedment (Figure 3.1) was proposed, and the calculation of the reference
embedment depth was done by equilibrium of horizontal forces and moments.
Figure 3.1 Engels method. Parabolic earth net pressure distribution
A large number of limit equilibrium design methods were developed along similar lines
(see Annex II), but differing with respect to:
-
The shape of the earth pressure distributions

The method of computing the equilibrium of the wall
The theories used in derivation of earth pressure coefficients
The definition and magnitude of a factor of safety
The assumed distribution and direction of wall friction
The method of obtaining the internal friction angle of the soil.
Kreys method
Krey (1932) proposed a rectilinear net earth pressure distribution (Figure 3.2) where the
active earth pressure in the back of the wall above the dredge line and passive earth
pressure in front of the wall immediately below the dredge line were fully mobilized
18
even before failure (Bica & Clayton, 1989), which was showed later by many other
authors. The design depth of penetration was calculated by finding z, corresponding to
the maximum net earth pressure in front of the wall, satisfying both equilibrium of
horizontal forces and moments about the bottom of the wall. Kreys method became the
basis of the most used method nowadays.
Figure 3.2 Kreys method. Rectilinear earth pressure distribution. This was originally
separated into two diagrams, for both upper and lower part of the wall from the dredge
level.
Blums method
A modification on Kreys method was proposed by Blum (1931), resulting in a
simplification of the previous methods. The earth pressure in the lower part of the wall
was replaced by a fictitious force R acting on point O, as shown in Figure 3.3. The
embedment depth was calculated by taking moments about point O. It was later when
the practice of increasing the calculated embedment depth by 0.2d was introduced (Bica
& Clayton, 1989), being the first Schmidt (1942). The final design embedment depth
was therefore 1,2d. Nonetheless, the safety provided by this practice was unknown
(Cuadrado, 2010).
Figure 3.3 Rectilinear earth net pressure distribution from Blums method. Note the
fictitious force R
19
Rowes method
Rowe (1951) carried out a comprehensive series of tests on model walls in loose and
dense sand. Starting from these, introduced an empirical design method. A factor of
reduction of 1.5 on the passive earth pressure and an angle of friction of the wall equal
to zero were considered in order to obtain the design depth of penetration, by taking
moments about the point of application of R. Equilibrium of horizontal forces was
verified as well, but Rowe did not consider (as Blum had done) the extension of the wall
below the point of application of R. To obtain the reference depth of penetration d, an
angle of wall friction equal to 2/3 was considered, and the passive earth pressure
remained unfactored. In fact, Rowe distinguished between the design embedment (no
wall-soil friction, reduction of Kp) and the equilibrium embedment (soil friction factored
2/3, Kp not reduced). The calculated depth of penetration was not increased by any
factor. The maximum bending moment was evaluated at the point of zero shear force
from a different net earth pressure distribution (shown in Figure 3.4 a), using the design
depth of penetration. However, now full mobilization was assumed of passive earth
pressures (calculated with an angle of wall friction of 2/3) immediately below the
dredge level and partial mobilization for depths below the depth to the maximum net
earth pressure t (Figure 3.4 b). An important feature in this method was an empirical
correction for the effect of wall flexibility on the maximum bending moment. Rowes
method was later simplified, by assuming a single net earth pressure distribution which
resembled Blums method, and by including factors of safety on shear strength
parameters when evaluating passive earth pressure.
(b)
Figure 3.4 Rectilinear net earth pressure distribution from Rowes method: (a) For
determining limit and design depths of penetration. (b) For bending moments at design
depths of penetration.
(a)
Brinch Hansens method

A more refined design method was introduced by Brinch Hansen in 1953. Unlike
previous methods, the Brinch Hansen method obtained a solution to the problem by
satisfying all plane equilibrium conditions, together with approximate kinematic
compatibility conditions between the wall and the rupture surface. This method was
extremely complex, and so an approximate step by step routine was recommended
which assumed a simplified earth pressure distribution (Figure 3.5). Brinch Hansen used
earth pressure coefficients derived on the basis of a factor of safety on c and tan, and
20
the position of the rotation centre relative to the depth at which the pressure was to be
calculated. The maximum bending moment was first determined, at the point of zero
shear force on this distribution. The depth of penetration was then obtained by requiring
equilibrium of horizontal forces and moments for the lower part of the wall, including
the unbalanced moment at the point of zero shear force. The additional depth of
penetration d was calculated using empirical equations. The lateral pressures were
assumed to be constant below the point of zero shear force; however Brinch Hansen
showed that this gave a solution within a few per cent of that achieved by the more
complex, full analysis.
Figure 3.5 Earth pressure distribution as assumed in Brinch Hansens method
Application of X-ray method

Bransby & Milligan (1975) presented an experimental study of the soil deformations in
laboratory-scale cantilever flexible sheet pile walls in dry sand. The wall was 30cm
high, embedded in 37 cm of dense sand, within a glass-sided tank. The study simulated
plane strain conditions. The X-ray technique was used to determine movements on the
soil around the wall and to observe the rupture surfaces in the sand. An analytical
method was presented, which linked soil deformations with deflexions of the wall.
The X-ray method was commonly used to measure deformations. In essence, the
method consists of burying a rectangular planar grid of lead markers within the sand
and the exposing radiographs of the sample at different stages of the test. The position
of each marker on the radiograph may be measured and so the displacements of the
marker, and hence of the soil, between exposures may be determined. An example of
results is shown on Figure 3.6.
21
Figure 3.6 Displacements in sand using the X-ray method in a test carried out by Bransby
& Milligan (1975)
The effect of wall roughness on deformation behaviour seems to be small. The method
proposed assumes that the increments of strain at points along any line (rupture
surface) are equal, and determined only by the incremental angle of rotation of the wall.
This method is found to give good predictions of the behaviour, and should make it
possible to estimate the magnitude and extent of surface settlements behind the wall
(Bransby & Milligan, 1975).
CIRIA Report 104

Padfield & Mair (1984) presented Design of retaining walls embedded in stiff clay, also
known as CIRIA Report 104, which has become one of the most referred publications
related to sheet pile walls. This report covers the design of singly propped or cantilever
walls embedded in stiff overconsolidated clay. An introduction to the earth pressure
calculation was included, and methods for both anchored and cantilever walls were
summarised (free-earth and fixed-earth conditions, full method and simplified method).
Various methods of assessing the factor of safety against overall stability were
compared, and four of them recommended. These were a factor on embedment, on
strength parameters, on moments and on passive pressure. Applied examples were
included as well.
Finite Element application
First applications of the Finite Element method date from early 1980s. The
discretizations of the geometry were coarse, and the results were limited.
Fourie & Potts (1989) presented a comparison of finite element and limit equilibrium
analysis for a cantilever wall. The soil behaviour was modelled through an elasto-plastic
constitutive law, under drained conditions. The classical methods used were the full
method and the simplified method, described in CIRIA Report 104. In the finite element
model, Youngs modulus is considered to increase linearly with depth. Drained
conditions, with zero pore water pressures everywhere, zero cohesion and a single soil
friction angle of 25 were considered. The domain was discretised by 165 eight-noded
isoparametric elements (see Figure 3.7). The effect of the initial stress was studied as
well. Two extreme cases, with K0=0.5 and K0=2 were performed.
22
Results showed that larger horizontal movement of the wall occurs with K0=2, than with
K0=0.5, for the same depth of excavation (see Figure 3.8). This figure shows a
horizontal asymptote, representing the maximum excavation depth allowed with this
geometry and these strength parameters, and therefore corresponds to the limit state.
The maximum excavation depth was 10.6m, with a wall height of 20m, which means an
embedment depth of 9.4m. Further results of the comparison indicated that both the
limit equilibrium method used in the Paper and the finite element approach give similar
predictions of the embedment depth required to ensure stability. The finite element
analysis justifies the 20% increase in embedment depth arbitrarily assumed in the limit
equilibrium approach, which was showed again by Cuadrado (2010). Results showed as
well, that in excavations with low K0 values, limit equilibrium methods overestimate the
maximum bending moments in almost 50%. Some reduction of the bending moment
must therefore be guaranteed.
(a)
(b)
Figure 3.7 (a) Finite element mesh performed by Fourie & Potts (1989). The squareshaped geometry has 100m side. (b) Displacements with K0=2.
Figure 3.8 Results of horizontal displacement vs. depth of excavation for two K0 values.
23
Day & Potts (1993) showed that while working with a Finite Element model, beam
elements are appropriate to represent sheet pile walls. However, since the effect of wall
element type is significant, beam elements are not recommended for the analyses of
thick concrete retaining walls (referring to rigid structures); in which 2-D elements are
appropriate as they more accurately describe the geometry of the structure.
Review of analytical methods
Bica & Clayton (1989) presented an in-depth review of the analytical methods available
until then. 25 methods were described and compared with each other and with
experimental data. The authors concluded that limit equilibrium methods may be
capable of providing good estimates of the equilibrium depth of embedment and
maximum bending moment. However, this will only occur if there are subtle
compensations for the errors inherent in their extremely simplified assumptions. The
authors mentioned as well the need to develop further research, primarily focused on
acquiring more experimental data on the behaviour of sheet pile walls.
Kings method
A new analytical method for the analysis and design of sheet pile walls at failure
conditions was proposed by King (1995). This method has become relatively popular
and is frequently referred. The method essentially consists of a semi-empirical
procedure, since it is based on limit equilibrium, and at the same time a parameter is
taken from a graph. It represents an alternative to the current limit state methods, not
requiring the geometrical simplifications of the earth pressure distribution. According to
this method, equilibrium can be established without assuming the limit state in the
passive region. On the other hand, the method is dependent on an empirical parameter
determined from the results of centrifuge tests. The applicability of this method has
been verified by Day (1999), proving its conservative prediction and proposing a new
value for the empirical parameter.
The active pressure is mobilised in the back of the wall above the dredge level, as
normally (see Figure 3.9). The rest of points on the graph can not be solved, although
they are related to each other. Starting from predetermined values of h and , y and x
(respectively, maximum pressure in front of the wall and depth of zero net pressure) can
be calculated. Below the dredge level, the slope of the net pressure law is unknown at
service conditions, but known at limit conditions. Then, x and P (active pressure at the
dredge level) may be related. The ratio x/P at failure conditions is known to be equal
to1/(Kp-Ka). By substituting P by Kah a ratio x/h at failure is obtained, and equal to
(Kp/Ka -1)-1. The safety factor is expressed as the ratio x/h at service conditions to (x/h)c
at failure conditions.
When designing for a predetermined safety factor, an embedment depth is calculated
following the procedure below. First of all, the ratio (x/h)c is calculated (at failure
conditions). It depends only on. Then, x/h is calculated at service conditions through
the safety factor. Finally, h/d is obtained from Figure 3.10 using =0.35 curve. Since h
is known from the start, the embedment depth d is obtained.
When checking the safety factor, the inverse procedure must be followed.
24
The recommendation to use =0.35 curve was based on centrifuge tests. Later, Day
(1999) showed that King's recommendation of =0.35 is generally conservative, more
than both the UK and the USA methods. This value must be therefore slightly lower.
Figure 3.9 Assumed net pressure diagram for Kings method
Figure 3.10 Obtention of the embedment depth d in Kings method. =0.35 curve must be
used.
King compares his method with the gradual method (called A in the paper) and the full
method (called B), and it is found to give critical depths of excavation closer to those
observed in centrifuge tests than the current methods. The main feature and benefit of
this method is the non-use of the currently assumed passive pressures in depth.
Powrie (1996) presented one of the few studies considering non-drained conditions.
Days method
Day (1999) presented a Finite Element study where the net earth pressure over the sheet
pile wall was examined. The mesh used (showed in Figure 3.11) presented some
improvements with respect to those used in previous papers. In the Finite Element Day
chose a wall height of 10 m and considered 5 cases with soil friction angles within 20
and 50, with variable excavation depth. Results are shown on Figure 3.12. The study
showed that the point of zero net earth pressure for the appropriate rectilinear
25
approximation is dependent on the ratio between the active and passive earth pressure
distributions at limiting conditions.
Figure 3.11 Finite Element mesh developed by Day (1999)
20
25
35
50
40
Figure 3.12 Cases studied by Day (1999)
Day proposed an equation in order to define the point of zero pressure (see Figure 3.13).
This equation proposed a linear relation between the position of the point of zero
pressure and the ratio Ka to Kp. Starting form this, and according to Day (1999), the
proposal by King (1995) that =0.35 is generally conservative. The predicted retained
height and maximum bending moment are generally less than indicated by the Finite
Element data. Results from the Days equation were found to be in excellent agreement
with the finite element predictions and with centrifuge model data.
26
Figure 3.13 Point of zero net earth pressure, presented by Day (1999)
On later studies, the same author explained that the value is known at all stages of
excavation but is not necessarily constant. Remember that is the quotient between the
distance from the toe of the wall and the point of zero net pressure and the embedment
depth d. The value of is given by the equation 3.1 (Day, 1999):
(3.1)
The value may be constant only under the assumptions of linear earth pressure
distribution, which is reasonably true on the back of the wall, but it is not in front of the
wall.
On a newer paper released in 2001, Day showed the results of 30 cases using the same
geometry and mesh, comprising a range of conditions with soil friction angles within
20 and 50 and K0 values of 0.5 and 2.0. The author notes that for practical cantilever
walls, the active and passive earth pressure coefficients can be assumed to be at the
theoretical limiting values of methods such as Caquot and Kerisel. Day (2001) verified
and noted the very good predictions give by the Padfield and Mairs CIRIA Report 104.
The maximum bending moment is dependent only in the net pressure above that point,
and so the maximum bending moment may be calculated directly using the assumption
that the passive pressure is fully mobilised. The expression of the maximum bending
moment may be therefore simplified as shown in 3.2.
(3.2)
Where,
(3.3)
Mobilizable Strength Design method

Osman and Bolton (2004) presented a new method for the design of cantilever sheet
pile walls by introducing the concept of mobilizable soil strength. The motivation was
the need to address the real and non-linear behaviour of the soil (Osman & Bolton,
27
2004). The objective was to satisfy both serviceability and safety in a single step of
calculation.
The concept of mobilizable strength refers to the shear stress in undrained conditions
of clays. Clays that remain undrained during the excavation stages must deform only in
shear. Figure 3.14 shows the idealised earth pressure distribution from which, starting
from a known geometry, strength and pivot position may be determined by solving
force and moment equilibrium.
An especially interesting tip in this study is the ratio of the height of the rotation point
above the toe and the total height of the wall, for different excavation depths. The graph
shows a maximum where the height of the rotation point is always less tan 2.5% of the
total height of the wall (see Figure 3.15).
Figure 3.14 Lateral earth pressure distribution under undrained conditions. Cmob refers to
the mobilizable soil strength (Osman & Bolton, 2004)
Figure 3.15 Height of rotation point above the toe. H = excavation depth, D = total wall
height. (Osman & Bolton, 2004)
Limit equilibrium methods do not exist in isolation. Each of the limit equilibrium design
methods was originally associated with particular earth pressure coefficients, methods
for determining angles of friction, safety factors, etc. These design facets are at least as
important as the fundamentals of the limit equilibrium calculations.
Bearing in mind that the equilibrium calculations most commonly used (those of Krey
and Blum) pre-date the widespread use of effective stress, of triaxial testing, and of the
standard penetration test, it is hardly surprising that a large element of adaptation has
occurred. The survey of the literature carried out by Bica & Clayton (1989) produced
about 25 different methods of design, and almost every one of those was based on either
28
Kreys or Blums methods. Since then, many other methods have been presented, such
as Kings method or Days method. The principal variations within these methods are in
earth pressure coefficients, factors of safety, shear strength parameters and the influence
of construction methods.
3.2.
Anchored sheet pile walls
Rowes experiments
In 1952, Rowe presented the results from fifteen models of flexible walls retaining
cohesionless soil (see Figure 3.16). The influence of surcharge, anchor yield, dredge
level, soil type and state, pile flexibility and anchor level were studied. Rowe presented
a chart for the design of anchored pile walls. One of the conclusions was that the anchor
load is generally close to the free earth support value, which is a common procedure for
the design of anchored sheet pile walls. The design procedure consisted on entering into
the chart with the soil friction angle and the parameter (see Fig.3.2.1). The last was the
quotient between the depth to the anchor level and thetotal wall height. The chart
provided the penetration depth indirectly, from the parameter , which was equal to
1
.
Figure 3.16 Anchored sheet pile wall modeled by Rowe (1952)
These design procedure was based on experimental data of reduced models. The anchor
inclination was horizontal. The different tie-rod levels and dredge levels provided
different curves .
29
Fig.3.2.1. Chart presented by Rowe (1952) for the design of anchored sheet pile walls
First studies of anchor inclination

In 1972, Plant presented a study of the anchor inclination on multi-anchored sheet pile
walls. The paper presented the results of tests with a scale-model sheet pile wall of 0,6m
height retaining dry sand (see Fig.3.2.2). The three anchor rods were inclined 0, 15,
30 and 45. In a second series, the two lower levels remained horizontal, while the
upper level was inclined successively 30, 45 and 60.
Fig.3.2.2. Sheet pile wall scale-model developed by Plant (1972)
Plant noted the complexity of the behaviour of the overall system soil-wall-anchor.
Plant showed that when increasing the anchor inclination, the wall and soil movements
were enlarged. Movement of the wall may not be eliminated but reduced by selecting an
30
appropriate anchor inclination. Additionally, having the lower level of anchors less
inclined than the upper level provides a more rigid integrated structural system that
reduces movements. In this publication, Plant noted the need of developing the Finite
Element method, which is one of the earliest mentioning to do so.
The main contribution of this study was to provide experimental data, which has a great
value, especially nowadays.
Radiographic method
Milligan (1983) presented an experimental study on sheet pile walls anchored near the
top where displacements were measured by the radiographic method, as done by
Bransby& Milligan in 1975 for cantilever sheet pile walls. The behaviour of the wall
was found to be, as expected, much more complex than in cantilever case. Model walls
for this study were 30 cm height retaining dry and dense sand. The dredge level was
changed successively, resembling a staged excavation. The roughness of the wall was
changed as well in different series.
As done in cantilever walls by the same author, the X-ray method consisted in burying a
grid within the sand and then exposing the sample to X-rays. The lead markers show the
movement of each point of the grid, as shown in Fig.3.2.3. Displacements measured by
the X-ray method in a staged excavation (Milligan, 1983). The location of failure
surfaces may be possible, where the contours of incremental shear strain show a narrow
zone of a very high strain gradient (see Fig.3.2.4).
Fig.3.2.3. Displacements measured by the X-ray method in a staged excavation (Milligan,

1983)
31
Fig.3.2.4. Location of failure surfaces (Milligan, 1983)
The study was focused on the measurement of the movement behind the wall, and not in
front of this. Studies of measurements of movements in front of the wall could not be
found for this study.
Propped walls
Richards & Powrie (1994) developed a study supported on the finite element method.
The mesh used can be seen in Figure 3.17. The objective was the study of the effects of
construction sequence and pre-excavation lateral earth pressures on the behaviour of a
sheet pile wall propped not anchored at both crest and dredge level. Results showed
minimum wall movements occurred with top down construction (single excavation
stage) with temporary props. Minimum bending moments occurred with construction in
open cut (see Figure 3.18). The study concludes that maximum bending moments and at
the same time minimum movements occur in the case of top down construction with
temporary props, because the temporary prop takes comparatively lesser compressive
load.
Figure 3.17 Mesh used by Richards & Powrie (1994)
32
This study included two cases of initial earth pressure coefficient: K=1 and K=2. The
effect of increasing the earth pressure coefficient was, according to Richards & Powrie
(1994) that the magnitudes of the reverse bending moments, which occur at later stages
towards the bottom of the wall, are increased.
Figure 3.18 Wall deflections for K0=1 with (a) two temporary props, (b) one temporary
prop and (c) no temporary props.
First publications related to sheet pile walls were essentially results from experimental
tests, normally compared with theoretical design methods. On a second period,
theoretical developments led to several new methods. Nowadays, is much more difficult
to find results from experimental studies on recent papers, as these are usually focused
on the Finite Element method. In the early 80s , with the development of numerical
methods, the experimental studies started a trend to disappear, as opposite to Finite
Element studies, which are somehow more effective or more accessible, in terms of
simplicity, and essentially, in invested time.
33
4. NUMERICAL MODEL
4.1.
Introduction
This chapter has the double aim of introducing a general scheme of the usage of the
commercial software used (Plaxis) and explaining the performance of the model used.
PLAXIS is a special purpose two-dimensional Finite Element computer program used
to perform deformation and stability analysis for many types of geotechnical
applications. Real situations may be modelled either by a plane strain or an
axisymmetric model. The user interface consists of four sub-programs (Input,
Calculations, Output and Curves) which are performed to be used successively in a
geotechnical modelling. These sub-programs can be treated one by one, forming a
logical and complete methodology in order to undertake a geotechnical problem.
The operation of each sub-program is briefly described below.
Input program
To carry out a finite element analysis using Plaxis, the user has to create a finite element
model and specify the material properties and boundary conditions. This is done in this
sub-program. The user must create a two-dimensional geometry model composed of
points, lines and other components, in the x-y plane. The generation of an appropriate
finite element mesh and the generation of properties and boundary conditions on an
element level is automatically performed by the Plaxis mesh generator based on the
input of the geometry model. Finally, the initial effective stresses are generated to set
the initial state. The procedure scheme is showed below.
General settings: Specify a filename, the model (axisymmetry or plane strain), the size
of the domain, type of element (6 or 15 nodes), units, etc.
Geometry: Creation of the geometry using points, lines, plates, geogrids, interfaces or
anchors.
Loads and boundary conditions: Specify distributed or point loads, prescribed
displacements, drains, wells and fixities.
Material properties: Specify the general properties of the soil and structures. Soils can
be modelled either through a user-defined model or one of the predetermined (MohrCoulomb, soft soil, hardening soil, jointed rock). Properties are applied to each element
of the geometry.
Mesh: The mesh is the composition of finite elements in which the geometry is
discretised. The basic type of element is the 6-node or the 15-node triangular element.
Plaxis allows for a fully automatic generation of the mesh, resulting on an
unstructured mesh. This mesh may look disorderly, but the numerical performance of
these meshes is usually better than for regular (structured) meshes. Users may also
customise the finite element mesh in order to gain optimum performance.
The element used in this study is the 15-node triangular element, which provides an
interpolation of fourth-degree order for the displacements, with twelve points of
Gausss numerical integration where stresses are evaluated. The 15-node triangular
34
element provides better exactitude, although it leads to a major consumption of

memory, and lower values of efficiency.
Initial conditions: Initial water pressures are generated by drawing the water table.
Then, the initial effective stress field is generated by selecting the K0 coefficient for
each cluster, not taking into account of the structural elements.
Calculations program
One the geometry model has been created, the mesh has been generated and the initial
conditions have been created, it is necessary to define what kind of calculations must be
realised. The calculations program differentiates plastic calculations, consolidation, phireduction and dynamic. The project may be divided in phases in order to simulate
construction stages, for instance, the activation of a surcharge or a staged excavation.
Choosing plastic calculation, as used in all the cases studied, leads to non linear
equations which have to be solved in each calculation phase in a series of load steps. In
Plaxis, there are diverse proceedings to solve non linear plasticity problems, based all of
them on the selection of the load step. The selection of the correct configuration is
automatically done by the program, to ensure the optimal results. This proceeding is
controlled by a certain parameters that preserve the equilibrium between the exactitude,
efficiency and hardiness. However, these can be manually selected in order to closely
monitor the exactitude and the load steps.
Output program
In the Output program the user may easily view the results and elaborate results lists
from the values originated from the calculation process. The main results from a finite
element are the node displacements and the stresses, including the stresses against the
structural elements if present.
Results may be checked through shadings, contours, lines, arrows or other
representations. As said, results can be given in tables, with a structure that resemble to
Windows environment, so these can be easily exported to the main Office applications
(usually Excel) in order to handle and study them appropriately.
Curves program
The Curves program contains all facilities to generate load-displacement curves, stressstrain diagrams or stress paths in pre-selected points of the geometry.
This sub-program has been not used in this study, inasmuch as the diagrams provided
are limited to previously selected points, and taking into account that results have been
fairly handled, it has proved not to be a useful tool for the development of this study.
4.2.
Modelling sheet pile walls
Modelling sheet pile walls in a finite element program as Plaxis involves a large amount
of parameters, including those for the soil, for structures or for loads, all necessary to
complete a calculation.
35
This chapter is devoted to the explanation of all geometrical and geotechnical data
adopted for the design of sheet pile walls used in this study.
4.2.1. Geometry
The first step consists of the creation of the geometry of the model. On the general
settings window, shown in Figure 4.1, the user can select the general model used (plane
strain), the element type (15-node triangle) and the dimensions. The size of the
geometry depends on the case studied (cantilever or anchored walls), the excavation
height and the embedment depth. In each case the influence of the excavation may be
different. Two considerations related to the geometry size must be taken into account:
-
Too large size of geometry leads to a major computational cost. Besides, the results
given on distant points may be not relevant, since these are out of the influence
range of the excavation.
Too small size of geometry may lead to wrong results, as these are under the
influence of the boundary conditions.
It has been proved that the same sheet pile wall modelled with a different geometry size
gives different results from each other. In order to review and verify previous results, an
equal geometry has been used as in the previous study, when running the cases for a
second time.
Figure 4.1 General settings window
Modelling the ground

The ground is represented by a set of independent lines. When a set of lines form a
closed polygonal an independent cluster is created, where the soil properties can be
applied. Diverse clusters can be created in order to model, for instance a stratified soil
or the ground layers that will be successively excavated in a staged progress.
Modelling sheet pile walls and interfaces
Sheet pile walls may be modelled through the element plate. Plates are geometrical
lines as well. Consequently when placing a plate, a back and a front of the wall are
quickly created.
Plates represent structural elements, resembling to a beam, and therefore a bending
stiffness (EI) and an axial stiffness (EI) must be introduced. The equivalent plate
thickness is calculated through the following transformation:
36
(4.1)
The usage of the element plate avoids using a dense discretization through the mesh for
modelling the sheet pile wall.
Alongside the wall, the element interface has been used with a double aim. The
interface pretends to eliminate the stress and deformation peaks produced in the corner
of a structure. At the same time, the soil structure interaction is simulated by decreasing
the soil resistance in this contact. Each interface has assigned to it a virtual thickness,
which is an imaginary dimension used to define the material properties of the interface.
The higher the virtual thickness is, the more elastic deformations are generated.
Modelling ground anchors
Geogrids may be used in combination with node-to-node elements to simulate a ground
anchor. The geogrid is used to model the grout body and the node-to-node anchor is
used to model the anchor rod (free length).
The anchor rod, which crosses the failure wedge, is usually made up of cord or steel
cable whose function is to transmit from the anchor point until the bulb, situated outside
the failure wedge, the stresses supported by the wall.
The anchor bulb consists of the cord enclosed by a grout body, usually of cement slurry,
due to its length and diameter (depending on terrain and load). The forces transmitted
by the anchor rod, and hence the earth thrust, are resisted by friction.
Plaxis element "node-to-node anchor", which simulates the free length of the anchor, is
an elastic element that allows the transmission of load from one to another element,
simulating the free length of the anchor. The node-to-node anchor consists of two points
where the stress is concentrated, since the intermediate length that separates only acts as
a transmitter of loads, and therefore the bulb is not well represented. Due to this, the
grout body has been simulated using the element "geogrid".
The geogrid fixed to the end of the node-to-node anchor can represent a linear
element on which to distribute the loads transferred by the free length. The diameter of
the bulb may not be simulated by the mechanism described. However, a good
approximation has been found by assigning a value of stiffness to the
geotextile equivalent to that of steel cable. Because this simulation, two important
phenomena in the design of anchors can not be reproduced: slippage bulb-cord and
slippage bulb-ground. Both checks should be carried out afterwards.
37
Figure 4.2 Geometry of the model. (1) Clusters. (2) Sheet pile wall. (3) Anchor rod. (4)
Grout body
4.2.2. Loads and boundary conditions

Plaxis allows entering into the model boundary conditions and known loads. Of all the
options available, only the "standard fixities" option has been used . This boundary
condition is used in many geotechnical problems, as a quick and comfortable
application. In the case of sheet pile walls it has been very useful because it fits the
needs of the analysis in the following ways:
- To the vertical lines corresponding to the X coordinate that constitute a boundary of
the model, a horizontal fixity is assigned in which the lateral displacement is zero
(Ux = 0).
- To the geometric lines that constitute the lower boundary of the model, a zero
displacement in both vertical and horizontal axis is assigned (Ux = Uy = 0).
Plaxis external loads are of two types: point and distributed. Loads must be activated in
the calculations program only when desired.
4.2.3. Material properties

The application of the constitutive equations of finite element model requires the use of
a number of parameters or previous data, depending on the type of material, so that the
problem can be solved. These soil properties are usually obtained from geotechnical
studies.
The computational power of finite element management is accompanied by several
variables that need a numeric value. The main difficulty regarding these variables lies in
knowing the precise meaning of each of them, often invisible in the classical analytical
methods, and taking magnitudes that are coherent with each other to provide
consistency to the model.
This thesis is the continuation of a previous thesis. The contribution of this master thesis
has been to study the influence of some parameters whose influence was previously
missed. The analysis has been focused on observing changes in the behaviour of sheet
pile walls when those parameters or most significant data are changed. That would
allow making the comparison with analytical calculation methods.
38
The main parameters that have been analysed are: friction angle, cohesion, rigidity of
sheet pile walls and anchors, and anchor tensioning force. Before addressing the
calculation methodology adopted it is useful to know the input value required by the
program for each material.
Material sets for soil
Most of the material data has been chosen to be identical as in the previous thesis, so
that the study to be relevant and comparable. Material properties are introduced in
Plaxis within three categories:
1) General (see Figure 4.3):
Material model: The material model used in all cases has been Mohr-Coulomb, a onedimensional approximation to the real behaviour of the soil.
Material type: In all cases a drained material has been selected, so the soil friction angle
and the cohesion are involved.
Specific weight (unsat): 20 kN/m3. The saturated specific weight has been considered to
be the same, as it is not relevant in a drained study.
The pore index and the permeability are not relevant, since it is a drained study.
Figure 4.3 General window for material sets
2) Parameters (see Figure 4.4):

Young modulus (E): The value chosen represents an average value for soils,
1.3105kN/m2. This value is consistent with the stress and the paths expected for the
cases analysed. The influence of this parameter has been studied in the previous thesis,
and the findings showed that too large or too small values of E may cause numerical
problems.
Poissons coefficient (): A value of 0.3 has been adopted, which is consequently with
drained conditions. Usually this value is adopted for drained soils, when variations in
the volume are significant.
Cohesion (cref): For the cases analysed, a value of zero has been adopted. When
considered, a value of 10 kN/m2 has been adopted.
39
Soil friction angle ( in Plaxis, in the thesis): The main strength parameter of the soil.
Its value is variable in this study and constitutes the element in which this study is
focused.
Dilatancy angle (): Has not been taken into account in this study, because its usage is
limited to strongly overconsolidated soils.
Alternative stiffness parameters: These parameters are automatically calculated from the
Young modulus and the Poissons ratio.
Advanced parameters: This option allows the elimination of tensile stresses (tension
cut-off), through which the soil has zero tensile strength. This option has been used in
all the cases studied and is perfectly in accordance with the behaviour of the sheet pile
walls studied.
Figure 4.4 Parameters window for material sets
3) Interfaces
Through this window the strength of the interface is selected. The rigid option has
been selected, in order not to decrease the resistance of the soil between the wall and the
ground. Thereby, the same hypothesis as in the classical methods is fulfilled.
Once the material sets are competed, the user must apply the material properties to each
cluster. The at-rest coefficient is manually introduced in the phase of initial stress
generation (see chapter 4.2.4).
Material sets for plates (sheet pile walls)
The data for plates is summarised in one window (see Figure 4.5).
Material type: The constitutive model which governs the relation stress-strain has been
selected elastic for all the cases analysed. This selection simplifies the calculations
and is in accordance with the reality, since sheet pile walls are designed to have an
elastic response.
Axial and bending stiffness (EA), (EI): An elastic modulus of 35106 kN/m2,
characteristic of a current concrete used in sheet pile walls, and an area and an inertia
corresponding to a thickness of 40 cm have been considered.
Weight (w): Is obtained by multiplying the unit weight of the plate material by the
thickness of the plate. Therefore, it has units of force per unit area.
40
Poissons coefficient (): For thin and flexible structures a value of zero is
recommended for the direction normal to the plane.
Figure 4.5 Window of plate properties
Material sets for geogrids (anchor bulb)

Geogrids represent geotextile elastic material extending in the direction perpendicular to
the plane and are only able to resist tensile stress. Its elastic behaviour is modelled by
the axial stiffness value:
EA=3.3104 kN/m
This value is calculated by dividing the node-to-node anchor value by the distance
between anchors, to provide an equivalent stiffness to the bulb.
Material sets for node-to-node anchors (anchor rod)
The proceeding for this tool is identical to the geogrid (see Figure 4.6). An elastic
behaviour without plastification has been specified (for the same reasons given in the
case of the plates). The axial stiffness value is the same as that used in the geogrid and
corresponding to a steel cable obtained form the design of the anchor.
The difference with the geogrid is the introduction of the spacing of anchors to
automatically obtain the equivalent stiffness.
Figure 4.6 Window for node-to-node anchor properties
4.2.4. Mesh and initial stresses generation

The division of the finite element domain in which to perform the subsequent
calculations is the discretization of the problem and is represented by the mesh.
41
The mesh has been generated by a preliminary analysis that uses a coarse mesh (few
elements). Then, points or areas with concentrations of stress are found, or strain or
phenomena where better accuracy is desired. The global mesh of the model in these
areas is adjusted, to take into account possible local phenomena, maintaining a certain
balance with the time consuming calculation.
It has been found that despite a very fine mesh provides better accuracy of results, the
consumption of computing time increases and therefore a decision has been made to
sacrifice accuracy to achieve a better balance over time.
The elements requiring special attention such as sheet pile walls or anchors, which have
demanded greater precision of the results for further analysis, the mesh is refined by
local refinement factor that reduces the size of the elements in proportion to the value
introduced. In general, the overall level of refinement type is "fine" with a local
refinement in displays and anchors between 0.25 and 0.5.
Having established the optimal mesh, the last step before proceeding with the
calculations is to establish the initial conditions for the water flow and stresses.
When setting the initial stresses the program switch off by default all the different
elements to the soil type, as a starting point of the tensions of the soil prior to any
action. At all times it has been assumed normally consolidated soil represented by a
zero value of OCR and POP parameters corresponding to the degree of
overconsolidation and preconsolidation pressure respectively.
The at-rest coefficient may be introduced manually (see Figure 4.7), and then assigned
to each cluster. Whether not, the program automatically introduces a value of the
coefficient calculated from Jackys formula by default:
After this step, the initial stresses are generated in the model (see Figure 4.8).
Figure 4.7 Window for entering K0
42
(4.2)
Figure 4.8 Initial effective stresses
43
5. INFLUENCE OF THE INITIAL STRESS STATE

The initial stresses in a soil body depend on the weight of the material and the history of
its formation. At the initial moment, the stress state is characterised by a vertical
effective stress (v0). The horizontal earth pressure is determined by the lateral earth
pressure (or at-rest) coefficient (K0).
The influence of initial stress conditions has been studied through a finite element
model by Fourie & Potts (1989). In this study, the geometry of 100X100 m was
discretised by 165 eight-noded isoparametric elements. Two values of K0, 0.5 and 2,
considered to be extreme, were used and a single soil friction angle (=25) was
considered. Nowadays, the development of finite element programs together with the
improvement of calculation power of computers, allows extending the range of cases
studied fairly easily.
The at-rest pressure is the initial thrust exerted by a horizontal-surface soil body before
being excavated; or before a wall attached to it suffers any kind of movement. Whether
the soil is homogeneous and the forces applied over it vertical or horizontal, this thrust
will be, by symmetry, horizontal. The at-rest coefficient is therefore defined as the ratio
horizontal to vertical effective earth pressure, as in eq. 5.1.
(5.1)
The finite element program Plaxis has an automatic procedure to introduce the value of
the at-rest coefficient (K0) when generating the initial stresses. By default, the parameter
K0 is calculated through the formula of Jacky (eq. 2.1). However, Plaxis allows
introducing manually it instead. A consideration must be done in order to ensure that the
soil is initially normally consolidated, since the formula of Jacky is appropriate only for
normally consolidated soils. The possibility of having an overconsolidated soil involves
the manual introduction of K0 using eq. 5.2 or eq. 5.3 (Mayne & Kulhawy).
(5.2)
(5.3)
In normally consolidated soils, K0 is constant and depends on the soil friction angle. In
overconsolidated soils, K0 is found to be much more variable with the variation of the
soil friction angle. The methodology used in this study, consisting of a phi-reduction
procedure, involves a different value of K0 when initial stresses are generated, since the
soil friction angle is different in each case. In order to make the results comparable, in
the previous work of Cuadrado (2010) a criterion was established, consisting on
generating the same initial stresses for all cases. That means introducing the same value
of K0 in the entire domain regardless of the soil friction angle with the purpose, as said,
to establish a basis in which results may be comparable.
This study is focused on the influence of the initial K0 value on the behaviour of the
sheet pile wall. Results of the failure and the stress state for the same geometry were
presented and analysed in the previous work of Cuadrado (2010). For this reason,
results are not shown for a second time.
44
When introducing manually the value of K0, the program Plaxis warns of selecting very
high or very low K0-values, since this values may cause initial plasticity. This effect will
be explained forward.
5.1.
Methodology
A phi-reduction procedure has been applied to two free embedded cantilever sheet pile
walls. The procedure resembles to that developed by Cuadrado (2010), consisting of
modelling the sheet pile wall on Plaxis and then calculating the equilibrium by reducing
the soil friction angle successively until equilibrium is not reached. The soil friction
angle corresponding to the limit state is assumed to be the last case immediately before
collapse.
Near the limit state, numerical difficulties are found out, hindering the choice of the
minimum soil friction angle supported. Eventually, equilibrium is achieved for a soil
friction angle lower than in a case where it is not. Due to this, a criterion of choice has
been established, based on selecting the first angle for which the equilibrium is not
fulfilled, in a phi-reduction procedure. In most cases, the numerical difficulties are
found to be related to the instantaneous excavation, which are, at least partially, solved
with a staged excavation.
The methodology followed in the study of the influence of K0 is presented below.
1) Model a cantilever sheet pile walls in Plaxis with a determined geometry:
h=d=5m.
2) After pre-selecting a value of K0, the limit state is found by reducing
successively the soil friction angle (phi-reduction procedure) as explained
above.
3) A new value of K0 is adopted, and point 2 is repeated. The range of K0
values used attempt to be realistic, within 0.5 and 2.
4) Interpretation and analysis of the results.
Cases considered
As said, the sheet pile wall modelled has a retained height of 5m and an embedment
depth of 5m.
The phi-reduction procedure starts from a soil friction angle of 45. The methodology
based on assuming a K0 value independently of the soil friction angle presents a
difficulty. This is a synthetic situation, and therefore verification must be done in
order to ensure the value of K0 adopted to be within the values of Ka and Kp, according
to relation 2.2. It should be not forgotten that Ka and Kp represent limit values, and
therefore K0 can not be lower as Ka or higher as Kp. In that case, the soil would be at
failure in the initial state. Hence, the possible values of K0 comprise a range as showed
on Table 5.1. Taking into account that Ka increases as decreases, and Kp decreases
with (eq. 2.8 and 2.9), the soil friction angle considered in Table 5.1 is expected to be
the minimum achieved in Plaxis for this geometry. According to this, K0 may not be
lower than 0.44 or higher than 2.28, which is in agreement with Fourie & Potts (1989),
who considered two extreme values, 0.5 and 2. Thus, four values of K0 have been
selected for the study: 0.5, 0.75, 1, 1.5 and 2.
45
Ka
Kp
23
0.44
2.28
Table 5.1 Selection of the cases to study
5.2.
Analytical methods
The sheet pile wall design usually starts from soil geotechnical data, basically the soil
friction angle. Then, the embedment depth is calculated according to the classical
analytical methods, seen in 2.4.1.
In this study, as in the previous work of Cuadrado (2010), the procedure has been the
inverse as the common way to design a sheet pile wall. In other words, the way to
proceed has been to start from a specified geometry (excavation and penetration depth)
and then obtain the soil friction angle corresponding to the limit state of stability. The
reason of this procedure is essentially practical. While working with the numerical
model, a huge amount of cases are needed to be run, and it is much easier to change the
soil properties than the geometry.
Table 5.2 shows the results for the case analysed.
ANALYTICAL METHOD
s
g
h=5m d=5m
28.5
30.1
Table 5.2 Soil friction angles resulting form the analytical methods. s=simplified method.
g=gradual method.
CASE
5.3.
Results analysis
A sheet pile wall with a retained height (h) of 5m and an embedment depth (d) of 5m
has been modelled. In this chapter the results of the behaviour of the wall related to K0
variation are explained in depth.
Movements of the sheet pile wall are mainly horizontal (see Figure 5.1). A vertical
component almost always exists, which can be ascendant or descendant. However, the
most part of the movement corresponds to a horizontal component. A rotation point is
recognised near the bottom of the wall, such as classical methods predict.
The horizontal displacement at the top of the wall has been selected in order to assess
the effect of K0 parameter on the wall behaviour.
Figure 5.1 Displacements in the sheet pile wall, represented with arrows, near the limit
state. A rotation point is recognised near the bottom of the wall.
46
Figure 5.2 shows the evolution of the horizontal movement at the top of the wall as the
soil friction angle is decreased. This relation does not represent a physical process or
evolution in time, but a set of results for the same geometry when the soil has
successively poorer strength properties. At first sight, a non linear trend is recognised.
The horizontal movement of the wall grows as the soil friction angle is reduced. Notice
the vertical asymptote corresponding to the limit state, with strong similarity to results
of Cuadrado (2010). Failure situations have not been represented.
K0= 0.3 case has been represented as well (see Figure 5.2), but the sequence of phireduction has not been completed, since the soil is at failure conditions for values of
below 35. This case is therefore not representative.
Curves are in agreement with each other. These are nearly parallel and have the same
asymptote, with the exception of K0= 0.5 curve, which seems to have another trend. For
K0 values from 0.75 to 2, movements at the top of the wall grow with K0, for the same
soil friction angle (see Figure 5.2). This observation is in accordance to those showed
by Potts & Fourie (1989). The difference on the wall horizontal displacement may be
considerable, being able to reach the double value of displacement depending on the K0.
A different trend can be recognised for K0 values below 0.75. In this case, horizontal
displacement seems to grow as K0 value decreases. Such fact involves the existence of a
minimum displacement for a particular K0 value.
130
120
110
100
Ux attop(mm)
90
80
Ko=0,5
70
Ko=0,75
60
Ko=1
Ko=1,5
50
Ko=2
40
Ko=0,3
30
20
10
0
15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45
()
Figure 5.2 Evolution of the movement at the top of the wall with soil friction angle
variation. Notice the vertical asymptote corresponding to the limit sate.
47
Figure 5.3 shows the horizontal displacement at the top of the wall, which is equivalent
to a vertical cut on Figure 5.2. As seen above, a minimum value of horizontal
displacement can be recognised for a particular value of K0, especially near the limit
state. This minimum is more recognisable, the closer from the limit state the soil is. The
minimum movement is located for K0 values between 0.7 and 0.9 depending on the soil
friction angle. In a limit situation with a soil friction angle of 25 the minimum
movement occurs with a K0 value of 0.85. Movement may change up to 40% depending
on the initial stress state.
Horizontalmovementattop(mm)
In Figure 5.3, the right branch of the graph represents a rigid movement, with higher
values of horizontal stress in relation to vertical stress; while the left branch shows a
more deformable behaviour pattern.
70
25
60
50
40
30
30
35
20
10
0
0
0,25
0,5
0,75
1,25
1,5
1,75
2,25
K0
Figure 5.3 Evolution of the movement at the top of the wall with K0 for different soil
friction angles.
Numerical difficulties
Some numerical difficulties are found near the point of minimum displacement. When
calculating this case, the maximum number of iterations is reached and the accuracy
condition is not fulfilled.
48
6. MULTIPLE-ANCHORED SHEET PILE WALLS

6.1.
Introduction
This chapter is devoted to the study of sheet pile walls anchored with multiple rows of
anchors. The behaviour of single anchored sheet pile walls has been studied by
Cuadrado (2010). This chapter pretends to be the broadening of this study to further
anchor levels.
As seen in chapter 2.4.3, it can be stated that the complexity of the system soil-structure
grows as the number of anchors is increased. The complex interaction between wall and
soil own of the design of cantilever sheet pile wall by itself is increased by adding the
effect of the anchors, which interact with both the wall and soil. The design of this kind
of structures is based on the global experience. In fact, there may not be found any
analytical method for the design of multiple-anchored sheet pile walls in the existing
literature. All the studies performed so far are essentially experimental.
The object of the study is three fold. First, to evaluate the influence of the construction
procedure in a two level-anchored sheet pile wall. Secondly, to evaluate the influence of
the variation of the anchor load, and the variability of the bending moments depending
on the moment in which the pre-stressing force is applied. And finally, a case of threelevel-anchored sheet pile wall has been developed, in order to compare it with the twolevel-anchored case and notice the advantages of a third row of anchors. These issues
have been studied deeply in the past by authors such as Richars & Powrie (1994) for the
case of propped not anchored sheet pile walls. A summary of this study can be found
on chapter 3.2.
The wall-soil system response may differ greatly depending on the construction
procedure. It is not the same applying the anchor load when the dredge level is at the
anchor point, or applying the same load after finishing the excavation, with the entire
sheet pile wall working as a cantilever structure. The bending moment law and the wall
movement may differ depending on the cases explained.
The anchors may not be placed before the excavation takes place. The excavation at the
anchor level must be achieved before placing the anchor and applying the pre-stressing
force. In fact, further excavation depth must be achieved before placing the anchor due
to constructive needs. Therefore, there is a lapse of time in which the sheet pile wall, or
a particular height of it, behaves such as cantilever wall.
In the same way, the anchor position may be subject to an arbitrary location. The
common practice is to locate the first row of anchors near the crest of the wall in order
to reduce movements at the top. Bending moments and movements may differ as well
when choosing the appropriate location.
The usual trend in the design is to reduce bending moments and movements. A lesser
bending moment may permit the reduction of the wall section, which represents an
economical and practical benefit. A reduced movement represents an important
advantage at service conditions. In anchor load terms, the predictable response to the
anchor load increase is to reduce the bending moment, which is the main motivation of
anchoring sheet pile walls. This is preferable to reducing the anchor load and increasing
the bending moment.
49
There are no geotechnical mechanisms for the failure of anchored sheet pile walls, but a
global failure mechanism. This may occur specially with anchors with insufficient
length, where the anchor does not overpass the Rankine failure wedge.
The inputs in this study are therefore the position and inclination of the anchor rows in
the wall, the construction procedure, obviously the geometry, and the anchor load. The
position of the anchor rows and the construction procedure are studied below. The
geometry is the same for all the cases. The anchor load is explained below.
Anchor load
The anchor load applied to the rod in a single-anchored sheet pile wall has been studied
by Cuadrado (2010). The anchor force of the tie rod may be determined by the
analytical methods free earth support or fixed earth support but this value
corresponds to the equilibrium at failure. However, this load will be different in service
conditions and depends on the stages followed.
Cuadrado (2010) showed as well that regardless the anchor force applied, after the
excavation is done, this trend to the theoretical value of the anchor force. In fact, the
fictitious point to which the anchor load tend is slightly higher as the theoretical free
earth support value. This behaviour may be seen on Figure 6.1.
450
400
Anchor load (kN)
350
300
250
200
150
100
50
0
0
10
12
excavation depth (m)

Figure 6.1 Variation of the anchor force as excavation takes place, for different
prestressing loads (Cuadrado, 2010).
Further conclusions from Cuadrado (2010) where that the behaviour of the wall
resembles to that predicted by the analytical methods only when applying to the anchor
a force similar to the theoretical value. As the strain takes place in the anchor, the
passive pressure is mobilized. In other cases with other values of prestressing force
the failure mechanism is different from that predicted by the classical methods and the
behaviour is not comparable.
The anchor load has been selected by basing on the free earth support method, and
therefore assuming a single row of anchors, a force of 249,01 kN/m must be applied to
the anchor in order to guarantee equilibrium. This force is supposed to be distributed in
50
both anchors. In order to simplify the cases, the same force is applied to each anchor
and is equal to the theoretical value for a single-level anchored sheet pile wall (124,5
kN/m). This value is rounded to 125 kN/m in order to guarantee the stability of the
model.
6.2.
Geometry and methodology
The geometry used consists of a sheet pile wall of 15 m total height and an embedment
depth of 6 m. The retained height above the dredge level is therefore 16 m. This
cantilever section is sustained by two rows of anchors with an inclination of 15. The
anchor length must be sufficient to overpass the Rankine failure wedge. Starting from a
strict soil friction angle of 36, an anchor length of 20 m fulfils this condition. The
geometry of the model may be seen on Figure 6.2.
Total lenght
15 m
Embedment depth
3m
Retained height
12 m
Anchor rows
2
Anchor inclination
15
Anchor length
20 m
Soil friction angle
36
Table 6.1 Sheet pile wall geometry
Figure 6.2 Geometry of the model for the double row-anchored sheet pile wall
6.2.1. Study of the influence of the construction procedure

For the study of the influence of the construction procedure a range of cases has been
considered. The aim is to recognise a trend in the relation between the stages followed,
the bending moment and wall movement. The methodology consists of the following
steps:
51
i) Represent the geometry of the model which permits a staged excavation.

ii) Installation of the anchors when the anchor level is reached in the excavation.
iii) Steeply increase the excavation depth before placing the anchors and applying
the load.
Cases considered
Case 1: Installation of the anchors when the anchor level is reached in the excavation.
Case 2 to 6: Late installation of the anchors with the maximum cantilever section as
possible on case 6.
CASES
CASE 1
CASE 2
CASE 3
CASE 4
CASE 5
CASE 6
Excavation before placing 1st row of

anchors
2m
17%
3m
25%
4m
33%
5m
42%
6m
50%
7m
58%
Table 6.1 Cases considered
Excavation before placing

2nd row of anchors
7m
8m
9m
10 m
10 m
11 m
6.2.2. Study of anchor load variation

In this study, the starting point is case 3 mentioned above. A case has been selected
where a certain wall defection occurs before placing the anchors. In this situation, an
excavation of 2 m below the anchor level is completed before placing the anchors.
The anchor load is steeply increased from 0.5 to 3 times the theoretical value (125
kN/m) as seen on Table 6.2. Therefore, the second case (125 kN/m) corresponds to the
case 3 mentioned above. The construction procedure is then the same as case 3, starting
with 4 m of excavation -2 m beneath the anchor level- and the placing and pre-stressing
of the upper anchor. Then the excavation continues until 2 m below the lower anchor.
The lower anchor is placed and pre-stressed and finally the rest of the excavation -3 mis completed. The construction procedure may be seen on Table 6.3.
Anchor load
0.5 x theoretical value
62.5 kN/m
1 x theoretical value
125 kN/m
1.5 x theoretical value
187.5 kN/m
250 kN/m
375 kN/m
Table 6.2 Anchor load variation cases
52
STAGE 1
STAGE 2
STAGE 3
STAGE 4
STAGE 5
Excavation 1 m
Excavation 1 m
Excavation 1 m
Excavation 1 m
Pre-stress upper anchor
STAGE 6
Excavation 1 m
STAGE 7
Excavation 1 m
STAGE 8
Excavation 1 m
STAGE 9
Excavation 1 m
STAGE 10
Excavation 1 m
STAGE 11
Pre-stress lower anchor
STAGE 12
Excavation 1 m
STAGE 13
Excavation 1 m
STAGE 14
Excavation 1 m
Table 6.3 Stages followed in the anchor load variation study
6.2.3. Three-level-anchored sheet pile wall

A case of sheet pile wall with three rows of anchors has been developed. The objective
is to notice the advantage of placing an additional anchor row in terms of wall
movement and bending moment. To do so, a third row of anchors has been placed
beneath the existing lower row, near the point with potentially higher values of bending
moment. Geometry may be seen on Figure 6.3. The expected behaviour is similar to the
double-row-anchored sheet pile wall case, but with lower values of the maximum
bending moment and horizontal wall movement. The construction stages followed are
summarized on Table 6.4.
STAGE 1
STAGE 2
STAGE 3
STAGE 4
STAGE 5
Excavation 1 m
Excavation 1 m
Excavation 1 m
Pre-stress 125 kN/m upper anchor
Excavation 1 m
STAGE 6
Excavation 1 m
STAGE 7
Excavation 1 m
STAGE 8
Excavation 1 m
STAGE 9
Excavation 1 m
STAGE 10 Pre-stress 125 kN/m middle anchor
STAGE 11
Excavation 1 m
STAGE 12
Excavation 1 m
STAGE 13
Excavation 1 m
STAGE 14
Pre-stress 125 kN/m lower anchor
STAGE 15
Excavation 1 m
Table 6.4 Stages followed in the 3-row-anchored sheet pile wall
53
Figure 6.3 Geometry of the model for the three-level-anchored sheet pile wall
6.3.
Results analysis
The wall horizontal displacement shows that all the length of the wall has moved to the
excavation (see Figure 6.4), and although the displacement is lower at the bottom of the
wall, the condition of the free earth support is fulfilled. The anchor load effect may be
seen on the stress diagrams, where the shape of the bending moment and shear curves
are modified with respect to the cantilever situation.
Figure 6.5 shows the final stress state for case 3, where the anchors have been placed
when the excavation has reached 2 m below the anchor point. It may be seen that the
passive pressure has been mobilised not only in front of the wall, but around the anchor
points. This situation is far from failure.
(a)
(b)
(c)
Figure 6.4 Sheet pile wall anchored in two rows with 200% of the theoretical anchor load.
(a) Horizontal wall displacement. (b) Shear diagram. (c) Bending moment diagram
54
Figure 6.5 Main stress directions around the wall for case 3
Influence of the construction procedure

Table 6.5 Results for construction procedure cases shows the results for the 6 cases
considered. The maximum horizontal displacement is minimum for the case 1 and
maximum for the case 6. There is a recognisable trend to increase with the lately
installation of the anchors (see Figure 6.6). This behaviour is not linear, but polynomial,
so the displacement is greater when the more excavation is completed before placing the
anchors. The effect of the early pre-stressing of the anchor is then to reduce the
movements, particularly at the top of the wall. Oppositely, a larger movement of the
wall takes place when lately applying the anchor pre-stress. A different behaviour may
be seen for the bottom of the wall (see Figure 6.7). This point has roughly the same
displacement independently from the stages followed. This behaviour is probably due to
an excessive embedment depth, greater than needed.
CASE %excv
1
2
3
4
5
6
17
25
33
42
50
58
MMAX
Displ.Hzt.MAX Displ.Hzt.TOP Displ.Hzt.TOE Mupperanchor Mloweranchor
(kNm/m)
(mm)
(mm)
(mm)
(kNm/m)
(kNm/m)
530,44
83,83
68
16,86
44
145,55
533,25
89,44
75,78
16,66
47,32
160,19
525,96
101,96
93,28
16,16
46,73
187,39
546,64
133,51
130,7
15,16
48,58
239,05
499,24
183,47
183,47
14,47
48,99
176,16
670,66
305,25
305,25
11,48
46,78
463,95
Table 6.5 Results for construction procedure cases
The maximum horizontal movement is not located in the top of the wall, but in a certain
depth. Additionally, the later the anchor force is applied, the closer the maximum
55
horizontal movement is to the top of the wall. This occurs for cases 5 and 6 (see Table
6.5). This behaviour is predictable, as it resembles to cantilever behaviour.
In a hypothetic extreme situation, where the 100% of the excavation is completed before
placing and pre-stressing the anchors, the entire wall is working as a cantilever structure
in the lapse of time dedicated to the excavation. And so the maximum displacement
would be at the top of the wall.
Max.hotizontalmovement(mm)
Maximumhorizontalmovement
350
300
250
200
150
100
50
0
0
10
20
30
40
50
60
70
%excavationbeforeanchorload
Figure 6.6 Maximum horizontal movement at the final stage for the 6 cases considered
Horizontalmovement
Horizontalmovement(mm)
350
300
250
200
Top
150
Toe
100
50
0
0
10
20
30
40
50
60
70
Figure 6.7 Horizontal displacement at the top and the bottom of the wall, at the final stage.
The shape of the bending moment distribution, and so the maximum bending moment,
depend on the anchor load. A clear trend has not been found for the bending moment
variation depending on the late application of the anchor load (see Figure 6.8). The
56
bending moment is supposed to be greater when the anchors are pre-stressed

immediately after reaching the excavation level and lesser when a lapse of time runs
before applying the load. This dependence has not been proved. The maximum bending
moment takes place at 10 m from the top of the wall, but this distance is smaller for
cases 5 and 6, where more deflection occurs before applying the anchor load. Note in
Figure 6.9 that the bending moment at the lower part of the wall has roughly the same
value, independently from the stages followed.
Maximumbendingmoment
Max.bendingmoment(kNm/m)
800
700
600
500
400
300
200
100
0
0
10
20
30
40
50
60
70
Figure 6.8 Maximum bending moment in the wall at the final stage
Bendingmoment(kNm/m)
Bendingmoment
500
450
400
350
300
250
200
150
100
50
0
Upperanchor
Loweranchor
10
20
30
40
50
60
70
Figure 6.9 Bending moment at the anchor points, upper and lower anchors
57
Anchor load variation

Five cases have been run with the same geometry, and only changing the anchor load
value. For the first case, a value of one half the theoretical anchor force has been
considered, and then increasing the load for the next cases until three times the
theoretical value in the last case.
Results on Table 6.6 show that the displacement decreases as the anchor load increases,
with a nearly linear trend (see Figure 6.10), which is the predicted behaviour. As
mentioned before, the top of the wall is more sensitive to the displacement variation
depending on the anchor load variation (see Figure 6.11). The upper row of anchors is
critical in order to reduce movements at the top.
%Overtheoretical Load
MMAX Displ.Hzt.MAX Displ.Hzt.TOP Displ.Hzt.TOE
(mm)
(mm)
(mm)
value(125kN/m) (kN/m) (kNm/m)
50%
62,5 539,12
119,13
115,25
15,58
Mupperanchor
(kNm/m)
47,56
Mloweranchor
(kNm/m)
224,56
100%
125,0
525,96
101,96
93,28
16,16
46,73
187,39
150%
187,5
518,35
87,62
74,42
16,80
48,53
141,63
200%
250,0
500,40
75,76
58,28
17,30
47,71
94,47
300%
375,0
448,28
52,83
37,91
18,34
Table 6.6 Results of anchor load variation
86,91
8,50
Maximumhorizontalmovement
140
120
100
80
60
40
20
0
0
50
100
150
200
250
300
350
%Anchorforcerespectthetheoreticalvalue
Figure 6.10 Maximum horizontal displacement while increasing anchor load
58
Horizontalmovement
140
120
100
80
60
Top
40
Toe
20
0
0
100
200
300
400
Figure 6.11 Horizontal displacement as anchor load is increased
In terms of bending moment, the expected behaviour is that a greater anchor load will
lead to greater values of bending moment. This pattern has not been seen on the cases
run. The bending moment, far from increasing, slightly decreases while the anchor load
is multiplied (see Figure 6.12).
Bendingmoment(kNm/m)
Maximumbendingmoment
600
500
400
300
200
100
0
0
50
100
150
200
250
300
350
Figure 6.12 Maximum bending moment of the wall for different anchor loads
Three-row-anchored sheet pile wall

The results obtained are the expected in terms of bending moment and horizontal
displacement. As seen in results Table 5.1 the maximum bending moment is
significantly reduced when adding a third row of anchors. In this table, results have
been compared with case 2 of the 2-row-anchored model, where the same excavation
stages have been followed. The modification on the bending moment distribution shape
may be seen on Figure 6.13. The effect of the lower anchor may be seen on the shear
59
distribution as well. In the same way, the maximum displacement is significantly

reduced (see Table 6.7). Its location depends, as in the 2-row-anchored model, on the
stages followed. When the anchors are placed after the whole excavation or most of it
has been completed, the structure has been working as a cantilever wall in this lapse of
time, and so the maximum displacement is registered at the top. The shape of the
horizontal displacement does not differ much from the 2-row-anchored case.
The anchor load has been considered the same for the three levels. However, it seems
that this practice may not be fully recommended. Anchor load may be adapted to the
shear stress distribution, applying higher load to the lower rows and a lower load value
to the upper anchor.
3rows
2rows(Case2)
MMAX Displ.Hzt.MAX Displ.Hzt.TOP Displ.Hzt.TOE Mupperanchor Mmiddleanchor Mloweranchor

(kNm/m)
(mm)
(mm)
(mm)
(kNm/m)
(kNm/m)
(kNm/m)
305,57
70,06
60,05
17,49
38,87
137,65
257,52
533,25
89,44
75,78
16,66
47,32
160,19
Table 6.7 Comparison between results for 3-row-anchored and 2-row-anchored sheet pile
walls
(a)
(b)
Figure 6.13 Comparison of diagrams, from left lo right, of horizontal displacement, shear
and bending moment. (a) 3-row-anchored sheet pile wall. (b) 2-row-anchored sheet pile
wall
60
61
7. CONCLUSIONS
The main conclusions of the study accomplished in this master thesis are presented
below.
Bibliographic review
First publications related to sheet pile walls date from the early 20th century. These were
essentially results from experimental tests, normally compared with theoretical design
methods. On a second period, theoretical developments led to several new semiempirical methods. Nowadays, after the relatively recent development of numerical
methods, it is much more difficult to find results from experimental studies in recent
papers. These are usually focused on the Finite Element method. In the early 80s, with
the development of numerical methods, the experimental studies started a trend to
disappear, oppositely to the Finite Element studies. These are somehow more effective
or merely more accessible in terms of simplicity and, essentially, in terms of invested
time.
Influence of the initial stress state K0
Depending on the initial stress state, the difference on the wall horizontal displacement
may be considerable, being able to reach the double value of displacement. This occurs
not only near the failure state, but at service conditions as well. For K0 values from 0.75
to 2, movements at the top of the wall grow with K0, for the same soil friction angle.
Oppositely, for K0 values below 0.75 the horizontal displacement seems to grow as K0
value decreases. This fact leads us to think that there is a particular K0 value for which
the displacement is minimal. This minimum is more recognisable, the closer from the
limit state the soil is, or in other words, with lower soil friction angle values.
For K0 values between 0.7 and 0.9 the minimum movement at the top is registered,
depending on the soil friction angle. With a soil friction angle of 25 (close to the limit
state) the minimum movement occurs with a K0 value of 0.85. Movement may change
up to 40% depending on the initial stress state.
Influence of the construction procedure
The construction procedure on two-level anchored sheet pile walls has a critical
influence over the movement of the wall, particularly at the top. It may be stated that the
later the anchors are placed and pre-stressed, the more deflection it occurs in the wall,
and so larger movements are registered at the top. Oppositely, when the anchors are
placed and pre-stressed immediately after reaching the anchor point, these absorb the
movement and limit the wall strains.
In contrast to cantilever sheet pile walls, the maximum horizontal displacement is not
located at the ground surface, but at a particular depth. However, it has been seen that
when applying the anchor force when most of the excavation has been completed, the
maximum displacement takes place at the top of the wall, instead of being at depth, and
resembling to the cantilever behaviour.
Influence of the anchor load variation
The anchor force increment has a direct effect on the reduction of the wall horizontal
displacement. When applying three times the theoretical anchor force, the maximum
62
displacement is reduced nearly times three. This reduction is particularly remarkable at

the top of the wall, although it obviously depends on the position of the upper anchor.
Oppositely, the lower section of the wall does not suffer this behaviour so closely. This
part of the wall slightly reduces its movement when increasing the anchor force.
Three-level-anchored sheet pile walls
The effect of adding a third row of anchors is to reduce movements and bending
moments. The upper row is preferable to be placed the near as possible to the top of the
wall. Its position is critical in order to reduce movements. The other rows may be placed
in arbitrary locations in order to reduce bending moment distribution.
It seems that applying the same load to both anchors is not the best practice. Anchor
load may be adapted to the shear distribution, applying higher load to the lower anchors
and less to the upper anchor.
This master thesis represents the continuation of the previous work of Cuadrado (2010).
The main objectives have been fulfilled: the state of knowledge has been reviewed,
cases have been run for a second time and new issues subject to review have been
discussed. An introduction to the modelization of multiple-anchored sheet pile walls
has been carried out and a scientific paper has been written in order to be published in a
scientific journal. However, this master thesis could include many other issues already
mentioned by Cuadrado (2010), such as un-drained conditions, presence of external
loads, etc. These have been excluded due to excessive extension and time, but may be
the subject of study in coming works.
63
8. REFERENCES
1.
BARDEN, L. (1974) Sheet pile wall design based on Rowes method. Part III of
A comparison of quay wall design methods. CIRIA Technical Note 54.
2.
BICA, A. V. D. & CLAYTON, C. R. I. (1989) Limit equilibrium design methods

for free embedded cantilever walls in granular materials. ICE Proceedings, Part 1,
86, 879-898.
3.
BLUM, H. (1931) Einspannungsverhltnisse bei Bohlwerken. W. Ernst & Sohn.

Berlin.
4.
BOWLES, J. E. (1988) Foundation analysis and design. McGraw-Hill, 4th

Edition. New York.
5.
BRANSBY, P. L. & MILLIGAN, G. W. E. (1975) Soil deformations near

cantilever sheet pile walls. Gotechnique 25, No. 2, 175-195.
6.
CHERUBINI, C., GARRASI, A., PETROLLA, C. (1992) The reliability of an

anchored sheet pile wall embedded in a cohesionless soil. Canadian Geotechnical
Journal, 29, 426-435.
7.
CHUGH, A. K. & LABUZ, J. F. (2011) Numerical simulation of an instrumented

cantilever wall. Canadian Geotechnical Journal, 48, 1303-1313.
8.
CUADRADO, A. (2010) Anlisis tenso-deformacional en rotura y condiciones de

seguridad de pantallas en voladizo y ancladas. Comparacin con mtodos clsicos.
Universitat Politcnica de Catalunya. Thesis.
9.
DAY, R. A. & POTTS, D. M. (1993) Modelling sheet pile retaining walls.

Computers and Geotechnics, 15, 125- 143.
10. DAY, R. A. (1999) Net pressure analysis of cantilever sheet pile walls.
Gotechnique, 49, No.2, 231-245.
11. DAY, R. A. (2001) Earth pressure on cantilver walls at design retained heights.
ICE Proceedings Geotechnical Engineering, 149, Issue 3, 167-176.
12. FOURIE, A. B. & POTTS, D. M. (1989) Comparison of a finite element and
limiting equilibrium analyses for an embedded cantilever retaining wall.
Gotechnique 39, No.2, 175-188.
13. HANNAH, T. H. (1987) Ground anchorages: anchor testing - measuring absolute
movement. ICE Proceedings, 82, Part 1, 639-644.
14. KING, G. J. W. (1995) Analysis of cantilever sheet-pile walls in cohesionless
soil. Journal of Geotechnical Engineering, Vol. 121, No.9.
64
15. MA, J., BERGGREN, B., BENGTSSON, P., STILLE, H., HINTZE, S. (2009)
Behaviour of anchored walls in soils overlying rock in Stockholm. International
Journal of Geoengineering Case Histories Vol. 2, Issue 1, 1-23.
16. MILLIGAN, G. W. E. (1983) Soil deformations near anchored sheet-pile walls.
17. OSMAN, A. S. & BOLTON, M. D. (2004) A new design method for retaining
walls in clay. Canadian Geotechnical Journal, 41, 451-466.
18. PADFIELD, C. J. &MAIR, R. J. (1984) Design of retaining walls embedded in
stiff clay. CIRIA Report 104.
19. PLANT, G. W. (1972) Anchor inclination - its effect on the performance of a
laboratory scale tied-back retaining wall. ICE Proceedings, Part 2.
20. Plaxis 8.2. (www.plaxis.nl)
21. POTTS, D. M. & FOURIE, A. B. (1984) The behaviour of a propped retaining
wall: results of a numerical experiment. Gotechnique 34, No.3, 383-404.
22. POWRIE, W. (1996) Limit equilibrium analysis of embedded retaining walls.
23. RICHARDS, D. J. & POWRIE, W. (1994) Finite element analysis of construction
sequences for propped retaining walls. ICE Proceedings, 107, 207-216.
24. ROWE, P. W. (1952) Anchored sheet-pile walls. ICE Proceedings. Part I, Vol. I.
25. ROWE, P. W. (1955) A theoretical and experimental analysis of sheet-pile walls.
ICE Proceedings. Part I, Vol. I.
26. ROWE, P. W. (1955) Sheet-pile walls encastr at the anchorage. ICE
Proceedings. Part I, Vol. I.
27. ROWE, P. W. (1956) Sheet-pile walls at failure. ICE Proceedings.
28. ROWE, P. W. (1957) Sheet-pile walls subject to line resistance above the
anchorage. ICE Proceedings.
29. ROWE, P. W. (1957) Sheet-pile walls in clay. ICE Proceedings. Part I, Vol. I.
30. SCHRIVER, A. B. &VALSANGKAR, A. J. (1996) Anchor rod forces and
maximum bending moments in sheet pile walls using the factored strength
approach. Canadian Geotechnical Journal, 33, 815-821.
31. SIMPSON, B. (2000) Partial factors: where to apply them? International
Workshop on Limit State Design in Geotechnical Engineering. Melbourne,
Australia.
65
32. KRABL, S. (2006) Interactional approach of cantilever pile walls analysis.

ActaGeotechnicaSlovenica, 2006/1.
33. USS (1975) Steel Sheet Piling Design Manual. United States Steel.
34. VALSANGKAR, A. J. &SCHRIVER, A. B. (1991) Partial and total factors of
safety in anchored sheet pile design. Canadian Geotechnical Journal, 28, 812-817.
35. YOUSSEF, Y. G. (2003) Berms for stabilizing earth retaining structures. Faculty
of Engineering of Cairo University. Fayoum, Egypt.
66
67
ANNEX I: PAPER
68
Safety assessment of limit equilibrium methods for the

design of cantilever and anchored sheet pile walls
Alejo Gonzlez, Agustn Cuadrado, Alejandro Josa, Sebasti Olivella
Abstract: The paper presents an accurate evaluation of the safety provided by the
classical methods based on limit equilibrium for cantilever and anchored sheet pile
walls. The most used of these methods are reviewed. The embedment depth increment
of 20% and the earth pressure coefficient reduction are established nowadays as
common practices, although the real safety provided by these is unknown. In order to
compare the prediction of the classical methods, the Finite Element program Plaxis has
been used. A decomposition of the safety factors has been developed in order to
evaluate the contribution of each practice to the global safety. Results show some
disagreements between the numerical solution and classical methods prediction. In
cantilever walls, the passive state is not fully mobilised at the bottom of the wall. In
anchored walls, the failure mechanism given by the Finite Element analysis suggests a
failure statewhere the sheet pile wall describes a translation movement on the deep
zone, and at the same time a rotation movement around the anchor point. The real safety
factor comprises a range between 1.3 and 1.8 for cantilever walls and between 1.3 and
1.7 for anchored walls, depending on the method and the anchor force. The safety factor
provided by the embedment increase of 20% is found to be 1.15 for cantilever walls,
and lower for anchored walls.
Key words: sheet pile wall, design, finite element, safety factor.
Introduction
Limit equilibrium methods in sheet pile wall design are being used since a long time,
even with the recent development of numerical methods, such as Finite Element
analysis. The full method for cantilever sheet pile walls, shown in Figure 0.1(a), has
been described by Padfield&Mair (1984), consisting on an idealised pressure
distribution at limiting conditions, and assuming that there is a sudden reversal between
the passive and active thrust at the rotation point. The complicated nature of this method
leads in practice to an iteration process, reason why the simplified version, shown in
Figure 0.1(b), is recommended (Padfield&Mair, 1984). A complete method, shown in
Figure 0.1(c), has been introduced by Bowles (1988), which considers a transition zone
around the rotation point where the thrust gradually changes its direction. Besides, a
large number of analysis and design methods for embedded cantilever walls have been
developed, being most of them based on classical limiting earth pressure distributions,
and these have been reviewed by Bica and Clayton (1986). The three mentioned
methods are well known as UK method, UK simplified method and USA method as
well (Day, 1999; krabl, 2006), according to the simplifications of the assumed earth
pressure distribution.
69
a) Full method or UK method
b) Simplified method or UK simplified method
c ) Gradual method or USA method

Figure 0.1. Limit equilibrium methods for cantilever sheet pile walls
70
The success of the extended use of the classical methods lies in its simplicity and its
proved reliability. Many authors have analysed these methods, noting that the collapse
of a cantilever sheet pile wall is in general well or slightly conservatively predicted
(Potts &Fourie, 1984; Powrie, 1996; Day, 1999; Day, 2001). The methodology used has
been diverse: centrifuge tests, laboratory-scale models, X-ray method, observations in
real walls, as well as analytical and numerical analyses. A semi-empirical method at
failure conditions has been proposed by King (1995): an alternative to the limit state
methods, not requiring the geometrical simplifications of the earth pressure distribution.
According to this method, equilibrium can be established without assuming the limit
state in the passive region. On the other hand, the method is dependent on an empirical
parameter determined from the results of centrifuge tests. The applicability of this
method has been verified by Day (1999), proving its conservative prediction and
proposing a new value for the empirical parameter. First comparisons between classical
methods and Finite Element showed that the actual net pressure distribution can be well
approximated through three lines (Day, 1999), which is in accordance with the methods
seen above.
Few authors have analysed the classical methods applied to anchored walls.
Valsangkar&Schriver (1991) showed the results of a parametric study, reviewed the
existing methods and proposed values of safety factors for each method. The X-ray
method has been commonly used to study experimentally the field of deformations near
anchored walls, and the complex phenomenon known as arching has been observed by
Milligan (1983). This refers to the state of stress existing in a zone of soil which is
trying to deform between two zones of more rigid soil (Milligan, 1983). The behaviour
of a sheet pile wall supported by multiple rows of anchors has been studied
experimentally by Plant (1972), who first mentioned the need to develop the finite
element analysis, and noted the complex behaviour of a multi-anchored retaining wall.
Finite element analysis applied to propped walls has been carried out by Potts &Fourie
(1984) and Richards &Powrie (1994), noting the first the relevance of K0, and observing
the second that the movements or the bending moments can be minimized by choosing
the appropriate construction sequence.
When designing through limit equilibrium methods, a factor of safety must be
introduced in order to take account of uncertainties, and to provide a safety margin.
Modes to achieve a safety factor are to increase the obtained depth of by 20%; to
decrease the two soil strength parameters c and ; to establisha relationship between
the restoring moments and the overturningmoments; as well as reducing the passive
earth-pressure coefficient (Padfield&Mair, 1984), (King, 1995), (Day, 1999). The first
practice is introduced as an attempt to take into account the simplification done when
modifying the earth pressure law (see Figure 0.3), although it may be used as well in
order to increase the global safety. It is found that the increase of 20% of the
embedment depthyields a conservative prediction (Padfield&Mair, 1984), although the
actual safety provided by the practices described is unknown.
The aim of this paper is to evaluate in detail the safety that these practices provide,
through an analysis of the safety factors. The limit state is reached in the Finite Element
program Plaxis by reducing successively the friction angle of the soil, represented by an
elasto-plastic law, and the contribution of each practice (Kp reduction and embedment
depth increment) is studied through a decomposition of the safety factors. The
71
simplified method and the gradual method are analysed in cantilever sheet pile walls,
and free earth support and fixed earth support methods are analysed in anchored sheet
pile walls.
Earth pressure distributions in cantilever walls

Stability in cantilever sheet pile walls depends on an adequate embedment below the
dredge line. The classical limit equilibrium methods attempt to model the sheet pile
walls at failure conditions, and differ from each other in several assumptions, but being
a common feature the reach of the failure state in the whole (or in the most part of the)
length of the wall.
Geotechnical design calculations in either cantilever or anchored sheet pile walls
establish equilibrium of horizontal forces and moments to define the failure state and
the reference embedment below the dredge line (before applying the safety factor).
Full method
This method, shown in Figure 0.2 and known as UK method as well, has been fully
described by Padfield&Mair (1984) and gets its name in contrast to the simplified
method, described later. The active limit state is assumed to be reached in the back of
the wall above the rotation point, and the passive limit state is assumed to be reached in
front of the wall between the dredge line and the rotation point. An overturn in the
normal pressure direction is supposed to be produced at the rotation point, below which
the full passive pressure is moved behind the wall and the active to the front, so there is
a sudden jump in the earth pressure distribution which is needed to prescribe moment
equilibrium.
Figure 0.2.Full method. Note the sudden change in the earth pressure distribution
Simplified method
Due to the complexity of the full method, a simplification is recommended
(Padfield&Mair, 1984). As shown in Figure 0.3, the earth pressure below the rotation
point can be replaced by an equivalent concentrated force acting on point O, represented
as R.
72
The value of d calculated is considerably lower than the one calculated from the full
method. The common practice is to increase it by 20%, being therefore the reference
embedment length equal to 1,2d. The zone below the rotation point is assumed to
receive the passive earth pressure, simplified as R, and a verification must be done in
order to ensure that the additional length 0,2d is able to generate the value of R.
The simplified method is slightly more conservative than the other methods, although it
leads to appreciably similar results(Padfield&Mair, 1984). Its greatest benefit is the
simplicity achieved on the traditional system of equations (
0,
0).
Figure 0.3.Simplified method. The thrust below the rotation

point O is replaced by a concentrated force R
Gradual method
A slightly different approach was made by Krey (1932), and later reviewed by Bowles
(1988), which does not involve the hypothesis of a sudden change in the earth pressure
distribution. The assumption made in this method is to consider a transition zone where
the net earth pressure gradually changes its direction from the front to the back of the
wall, as shown in Figure 0.4. This transition occurs around the rotation point and is
assumed to be linear. The earth pressure at the bottom of the wall is known (see Figure
0.4), so the incognita to obtain the embedment depth is the height z. The gradual method
is known as USA method as well.
73
Figure 0.4. Gradual method, shown in net pressures
Earth pressure distributions in anchored walls

Anchored walls with a single row of anchors are able to achieve the equilibrium without
considering a passive reaction at the bottom of the back of the wall. The tie rod position,
its inclination and its load are the subject of continuous studies, such as those carried
out by Plant (1972), Schriver&Valsangkar (1996) or Ma, Berggren, Bengtsson, Stille
and Hintze (2009).
Free earth support method

This method is based on the assumption that movements on the embedded zone of the
wall are sufficient to mobilize the active and passive thrust behind and in front of the
wall respectively. The passive pressure is assumed to act only in front of the wall
through the depth d (Figure 0.5). The bottom of the wall has therefore free movement,
and a minimum reference embedment depth, to satisfy equilibrium, is obtained.
Figure 0.5. Free earth support method. The equilibrium involves movement of the bottom
Fixed earth support method

Increasing the embedment depth, an overturn in the normal earth pressure is achieved,
until the deflected shape at the bottom becomes approximately vertical. Hence, the earth
pressure distribution is essentially equal to the cantilever casewith the simplified
distribution (see Figure 0.6), and the wall behaves like a partially built-in beam
subjected to bending moments (United States Steel, 1984).
74
Figure 0.6. Fixed earth support method (simplified distribution)
This method leads to a maximum reference embedment depth. It requires an additional

equation which is usually defined estimating where the zero net pressure point
(inflection point O; zero bending moment) is located. It is frequent to locate such point
in the middle of the embedment depth, though other assumptions have been proposed.
Blum (1931) showed a dependence relationship between the distance from the inflection
point to the dredge level and Ka.
Numerical analysis
In order to study the prediction of the classical methods, a Finite Element analysis has
been developed, assuming that its results are closer to the real behaviour of a sheet pile
wall. The software used was Plaxis 8.2., a 2-D Finite Element program. A homogeneous
single layer of dry soil has been modelled by an elastic perfectly-plastic law under plane
strain conditions. The element type is the 15-node triangle and the failure criterion is
Mohr-Coulomb.
The domain of the model is 75m to 90m length per 30m high, as seen in Figure 0.7 and
Figure 0.8. A sufficient size of the domain has been used, and it has been checked that
increasing its size the results remain unchanged. The boundary conditions are vertical
fixities on both sides of the geometry, and a total fixity on the bottom, reproducing the
most truthful field conditions. The specific weight of the soil is 17 kN/m3, its elastic
modulus 1.3105 kN/m2 and its Poisson coefficient 0.3, according to the drained
conditions. The soil is assumed to be cohesionless. When the effect of cohesion is
studied, the value considered is 10 kN/m2. The soil-structure interaction is modelled by
the element interface, which is assumed to be rigid. The soil resistance is therefore not
reduced between the soil and the wall. The K0 coefficient is assumed to be 0.5 in all the
cases studied. Further studies showed that for moderate values of K0 (0.4 to 1.7)results
do not differ significantly from those given for K0=0.5.
75
Figure 0.7. Finite Element mesh in a cantilever wall
Figure 0.8. Finite Element mesh in an anchored wall
The sheet pile wall is represented by an elastic element, which is a reasonable

approximation, as they are usually designed to have an elastic response. The elastic
modulus is 3.5107 kN/m2, the wall thickness 40 cm, leading to an axial stiffness of
1.4107kN/m and a bending stiffness of 1.867105kN/m. The specific weight of the
reinforced concrete is 28 kN/m3.
The grout body of the anchor is modelled through a geogrid element, a slender structure
with no bending stiffness, and an axial stiffness of 3.3104 kN/m. The anchor rod (free
length) is modelled through a node-to-node anchor with an axial stiffness of 5.8104 kN
and a spacing of 1.75 m. These values are commonly used for the design of sheet pile
walls and are summarized in Table 2.
Properties of the soil
Specific weight
Elastic modulus
Poisson coefficient
Cohesion
K0
Properties of the sheet pile wall
Specific weight
Elastic modulus
Thickness
Axial stiffness
Bending stiffness
Properties of the anchors
76
17 kN/m3
1.3105 kN/m2
0.3
0 kN/m2
0.5
28 kN/m2
3.5107 kN/m2
40 cm
1.4107kN/m
1.867105kN/m
Grout body axial stiffness

3.3104kN/m
Anchor rod axial stiffness
5.8104kN
Spacing
1.75 m
Table 2. Summary of soil, wall and anchor properties
While designing through classical methods, a value of the reference embedment depth d
is obtained to satisfy equilibrium, starting from a value of h and the soil strength
parameters. In the Finite Element analysis conducted, a different procedure has been
followed. The limit state is reached for each case studied by drawing the geometry, with
a known total wall length and embedment depth. The actual soil properties (including
soil friction angle) are considered to be those corresponding to the application of the
classical methods. The soil friction angle is then reduced successively until the collapse
of the soil body is achieved. This procedure enables the comparison between classical
methods and numerical methods, and assuming numerical methods to represent the
real behaviour allows evaluating the safety of the assumptions made in the classical
methods.
Bearing in mind that, in the numerical model, when collapse occurs, the soil-structure
equilibrium is not guaranteed, the solution must be an approximation to the limit state,
and this is assumed to be reached in the case just before collapse.
Near the limit state, numerical difficulties are found out, hindering the choice of the
minimum soil friction angle supported. Eventually, equilibrium is achieved for a soil
friction angle lower than in a case where it is not. Due to this, a criterion of choice has
been established, based on selecting the first angle for which the equilibrium is not
fulfilled, in a phi-reduction procedure. In most cases, the numerical difficulties are
found to be related to the instantaneous excavation, which are, at least partially, solved
with a staged excavation.
Cases considered
Six cases are studied, including cantilever and anchored cases. This represents an
adequate, to be meaningful, range of cases. Five cases are studied in cantilever walls
with excavation depths of 3 m and 5 m, and embedment depths within 3 m and 7.5 m. A
case considering the effect of cohesion is included. A single case of anchored wall is
studied, with 10 m of retained height and 2.5 m of embedment depth. Cases are
summarized in Table 3 below:
h:3m d:3m
h:3m d:4.5m
Cantilever sheet pile walls h:5m d:5m
h:5m d:7.5m
h:5m d:7.5m c:10kN/m2
Anchored sheet pile walls
h:10m d:2.5m
Table 3. Cases studied
Cantilever walls
77
In the first place, and starting form the established geometry, the soil friction angle is
calculated for each analytical method simplified and gradual methods considering all
the possible combinations of embedment increase and passive reduction, in order to
assess each effect separately. Consequently, 8 values of soil friction angle are obtained
from the analytical methods for each geometry. On the other hand, a phi-reduction
procedure through the Finite Element analysis is done with the same geometry. The
procedure leads to 2 values of soil friction angle, the first considering the original wall
geometry and the second considering the embedment increment. Table 4 summarizes
the terminology of the different angles calculated. Hence, 10 values of soil friction
angle are obtained; 4 from the simplified method, 4 from the gradual method and 2 from
the numerical method.
Results are given on Table 5. Note that P >P >g>s>FE. The effect of cohesion is
found to be, as expected, very beneficial for the stability.
20%
Embedment
increase
No
No
2/3 Passive
pressure
reduction
No
Yes
Simplified
method
Gradual
method
Yes
No
Yes
Yes
Numerical
method (
reduction)
Table 4. Soil friction angles obtained in cantilever walls

CASES
ANGLE OF SOIL FRICTION
28.5
30.1
25.2
26.7
33.5
35.0
30.3
31.7
25.0
22.0
4,5
21.4
22.8
18.7
19.9
26.7
28.0
24.1
25.3
18.0
16.0
28.5
30.1
25.2
26.7
33.5
35.0
30.3
31.7
26.5
23.4
7,5
21.4
22.8
18.6
19.9
26.7
28.0
24.1
25.3
21.0
18.0
8.0
7.0
5(c) 7.5(c)
11.9
10.3
16.6
15.4
Table 5. Cantilever walls. Soil friction angles in the limit state from the classical methods
and Finite Element. In the last row, (c) indicates cohesion of 10 kN/m2. Angles in degrees,
h and d in meters.
Anchored walls
As in cantilever walls, different combinations of the analytical methods with
embedment increase and passive reduction are considered. Consequently, for the same
geometry, eight soil friction angles are obtained from the classical methods and 4 from
the Finite Element analysis.
In the case of anchored walls, the variation of the soil friction angle supported
depending on the anchor force has been studied. The anchor force, known as the thrust
to be resisted by the anchor, is found starting from the free earth support method,
without passive pressure reduction. For the geometry considered, with a retained height
of 10 m and an embedment of 2.5 m, the anchor force resulting from the free earth
support method is 138 kN/m. This value, and its double amount, are considered in the
numerical model, and provide two different soil friction angles.
78
Table 6 summarizes the nomenclature of the different angles calculated. The two latter
represent 100% and 200% of the anchor force, in relation to that obtained from the free
earth support method with no passive earth pressure reduction.
Results are given on Table 7. The reduction of the passive earth pressure causes, as
expected, an increase of the soil friction angle with respect to the reference case. At the
same time, the embedment increase of 20% causes a decrease of the soil friction angle
with respect to the reference case. The variation of the anchor force, from 100% to
200% with respect to the theoretical value, provides a lower soil friction angle.
Numerical results afford, at first sight, lower values of the soil friction angle.
20%
Embedment
increase
2/3 Passive
pressure
reduction
Free earth
support
method
Fixed earth
support
method
No
No
fx
Yes
No
No
Yes
Yes
Yes
Numerical
method:
100% anchor
force
Numerical
method:
200% anchor
force
100
100
200
200
Table 6. Soil friction angles obtained in anchored walls

CASE
ANGLE OF SOIL FRICTION
fx
2.5
37.1
38.5
34.0
37.0
41.6
42.9
38.7
41.5
100
37.0
200
31.0
100
30.0
200
28.0
Table 7. Anchored walls. Soil friction angles in the limit state for the reference and design
embedment depth from the classical methods and Finite Element. Angles in degrees, d in
meters. Retained height h=10m
Results
Results: Cantilever walls
The stress state near the limit state described by the numerical method is broadly in
agreement with the predictions of the classical methods, as shown in Fig.10. However,
clear differences in the principal stresses are found between the numerical solution and
the classical methods, especially in the back of the wall around the bottom. The stress
state at this point should be theoretically close to the passive state. The numerical results
show that even though a turn in the principal stresses occurs, these are still relatively far
from the passive state (see Figure 0.9). It should be noted, that the classical methods
assume perfect plasticity and that the principal stresses can only be vertical or
horizontal. The numerical net pressure law differs as well in front of the wall
immediately below the dredge level, staying above the prediction of the classical
methods (see Figure 0.10 and Figure 0.11).This overpressure is due to the 3-D stress
state existent in Plaxis, which increases the confinement, so a greater resistance to
deformation is produced. The overpressure is higher the farther the soil friction angle
from the limit state value is. On the other hand, the active pressure provided by the
numerical solution matches the classical methods fairly well.
79
Figure 0.10 shows the evolution of the earth pressure distribution over the wall as the
excavation takes place. The limit state is restricted initially to the most superficial zone,
and the passive state in the bottom of the wall is reached in the last meters of
excavation. As shown by Kray (1932), Bransby& Milligan (1975), Powrie (1996), Day
(1999), and more authors, the active and passive pressures are fully mobilised above
and immediately below the dredge level respectively.
Retained height h (m)
Figure 0.9. Principal stresses represented with crosses near the limit state. The larger line
represents the major principal stress
Front of the wall
150
100
50
10,5
10
9,5
9
8,5
8
7,5
7
6,5
6
5,5
5
4,5
4
3,5
3
2,5
2
1,5
1
0,5
0
Stage1
Stage2
Stage3
Stage4
Stage5
Back of the wall
50
80
100
150
200
kN
250
Figure 0.10. Overlay of net earth pressure distributions near the limit state in a cantilever
wall, showing the evolution by excavation stages. h=5m d=5.5m =25
Top horizontal displacement (mm)
Figure 0.11 shows the displacement at the top of the wall growing as the soil friction
angle declines. This figure may be used to identify the soil friction angle at failure. A
vertical asymptote may be recognised, corresponding to the limit state. Figure 0.11
shows clearly a non linear trend on the displacement as approaching the limit state. The
evolution of the earth pressure lawas soil friction angle declines (Figure 0.12) shows
that the stresses in the lower zone of the wall are increased as the soil friction angle is
reduced.
80
70
60
50
40
30
20
10
0
20 22 24 26 28 30 32 34 36 38 40 42 44 46
Soil friction angle ()

Figure 0.11. Evolution of the horizontal displacement at the top of the wall with the soil
friction angle in a cantilever wall. Note the vertical asymptote
81
10
26,8
9
28
30
40
45
Height (m)
6
5
4
3
2
1
0
150
100
50
50
100
150
200
250
kN
Figure 0.12. Evolution of the net earth pressure law as soil friction angle is reduced in a
cantilever sheet pile wall with h=5m and d=5m
Results: Anchored walls

The analysis of anchored sheet pile walls reveals that the anchor affects the stress state
around the top and the bottom of the wall. As seen in Figure 0.1, a trend to generate a
passive state in the back of the wall around the anchor is found, as opposed to what is
established by the classical methods.
The numerical solution shows a rotation around the anchor point, which is in agreement
with the classical methods. However, a passive state in the back and bottom of the wall
is generated (see Figure 0.2), not satisfying the free earth support method hypothesis, as
it assumes the passive pressure only to act in front of the wall below the dredge level.
82
Figure 0.1. Principal stresses around the anchorage

12
11
Stage1
Stage2
Stage3
Stage4
Stage5
10
9
Height (m)
8
7
6
5
4
3
2
1
0
200
150
100
50
50
100
150
200 kN 250
Figure 0.2. Overlay of pressure laws as a staged excavation takes place in an anchored
sheet pile wall with h=10m and a final d=2,5m. =31 (near the limit sate)
For an equal soil friction angle, the fixed earth support method requires further
embedded depth than the free earth support method. Thereby, for an equal embedment
83
depth, the free earth support method provides a lower value of the soil friction angle as
the fixed earth support method. The Finite Element model does not consider the
conservative hypotheses of the classical methods, and therefore the limit equilibrium is
reached with a lower value of the soil friction angle. The existence of a threedimensional stress state in Plaxis lets the confinement grow, and therefore the resistance
to deformation is increased. The passive pressure is then greater than those calculated
from the analytical methods.
Anchor load results
Classical limit state methods provide an anchor force, while establishing equilibrium at
failure state. However, they neither consider the interaction with the soil, nor its
evolution in service, and the anchor load depends directly on the stages followed.
The anchor prestressing load has been changed from 0% to 300% of the theoretical
value. Figure 0.3. Evolution of the anchor load as excavation takes place, for different
prestressing loads summarizes the results and shows that the anchor load appears to
trend to an equilibrium value.
450
400
Anchor load (kN)
350
300
250
200
150
100
50
0
0
10
12
excavation depth (m)

Figure 0.3. Evolution of the anchor load as excavation takes place, for different
prestressing loads
Discussion
Cantilever walls
Results from the analytical methods are found to be in disagreement with the numerical
predictions. At first sight, the latter provides lower values of the soil friction angle, for
the same wall geometry. A comparison between methods can be established in terms of
safety factors, in order to asses the safety introduced by those.
84
Safety factors are typically expressed in terms of embedment depth, moments, passive
resistance or strength parameters (Padfield&Mair, 1984). Nonetheless, the value of the
factor obtained is not in accordance with the real safety provided. In service conditions,
a safety factor on the strength parameters can be defined as:
(4)
wherereal is the actual value of the soil friction angle and failure is the minimum value
supported, immediately before collapse. After increasing the embedment depth by 20%,
a lower value of the soil friction angle can be achieved before collapse. Hence, the
relationship between the supported friction angle before and after the increase of the
embedment depth gives an estimation of the distance to the limit state predicted by the
method due to this practice, as seen below:
Starting from (
(5)
(5), a first group of
safety factors (SF1 to SF5 on Table 8) reveals the safety provided by the increase of the
embedment depth by 20% for each method. Results show that this practice barely
provides a safety factor of 1.15, or lower.
The way to establish a more realistic safety factor must include the classical limit
equilibrium methods and the numerical methods. The safety factor can be decomposed
on successive multiplicative sub-factors that enable to rate each effect separately.An
accurate evaluation of the real safety of both simplified and gradual method can be
defined as follows:
(6)
Where tan refers to the minimum soil friction angle supported by the sheet pile wall,
calculated using the gradual method (tang) or the simplified method (tans). See SF6 to
SF9 on Table 9. Results show that the real safety provided by the classical methods
comprises a range between 1.2 and 1.8. The lower safety factor calculated corresponds
to the simplified method, while the greatest values are obtained for the gradual method
with passive earth pressure reduction.
The safety provided by the classical methods by ignoring the cohesionis known to be
favourable, and can be evaluated by adding the effect of cohesion on (
(6) as follows:
(7)
where the superscript C indicates cohesion. Table 10 shows the effect of ignoring the
cohesion in the simplified method. SF10 compares the soil friction angle calculated with
the simplified method versus the prediction of the Finite Element method considering
cohesion. SF11 on Table 10 further consider the reduction of the passive resistance.
Results show that ignoring, when existing, the cohesion turn out to be an especially
85
conservative practice. Otherwise, considering a wrong value of cohesion may have

severe consequences.
SAFETY FACTORS
SF1
tan
tan
tan
tan
SF4
SF3
SF2
tan
tan
SF5
tan
tan
tan
tan
1.16
1.16
1.13
1.13
1.15
h:3m d:4,5m
1.16
1.16
1.13
1.12
1.13
h:5m d:5m
1.16
1.15
1.13
1.13
1.15
h:5m d:7,5m
1.16
1.16
1.12
1.12
1.18
h:5m d:7.5m c:10kN/m2
1.16
CASE
h:3m d:3m
1.08
1.14
Table 8. Safety factors related to cantilever walls. Effect of increasing the reference
embedment depth by 20% for each method
SAFETY FACTORS
SF7
tan
SF8
tan
SF9
tan
tan
tan
tan
tan
h:3m d:3m
1.4
1.4
1.6
1.7
h:3m d:4,5m
1.4
1.5
1.8
1.8
h:5m d:5m
1.3
1.3
1.5
1.6
h:5m d:7,5m
1.2
1.3
1.6
1.6
h:5m d:7.5m c:10kN/m2
1.7
CASE
SF6
tan
2.4
Table 9. Safety factors related to cantilever walls. Safety provided by the classical
methods. Reduction of the passive pressure is evaluated for both methods non SF8 and
SF9
SAFETY FACTORS
SF11
SF10
CASE
tan
tan
h:5m d:7.5m c:10kN/m2
3.2
tan
tan
4.1
Table 10. Safety factors related to cantilever walls. Safety provided by the classical
methods when cohesion of 10 kN/m2 is ignored in addition
86
Anchored walls
The same relation as shown in (
(5)
is appropriate for anchored walls, and the values of SF12 to SF16 (see Table 11)
express the safety provided by the increase of the embedment depth by 20% for each
method. This practice leads to a safety factor of barely 1.1, depending on the method
and on the anchor force. Therefore, the embedment increase in anchored walls has less
influence on the safety than in cantilever walls.
The real safety of both free and fixed earth support method can be defined as in (3).
Safety factors SF17 to SF20, shown in Table 12, make out the situations of passive
earth reduction and two different anchor forces, in relation to the numerical prediction
including the increase of 20% on the reference embedment depth. Safety factors
comprise a range between 1.3 and 1.7, depending on the method and the anchor force.
As expected, safety factors grow from SF17 to SF20, and the fixed earth support
method with passive earth pressure reduction is found to be the most conservative. In
addition, note that an anchor force increase of 200% does not lead into a double safety
factor. The anchor force should trend to the equilibrium force when all the soil is at
failure, and so the safety factor considering 100% or 200% of the anchor force may not
differ greatly.
The analysis of the anchor force shows that when considering its theoretical value, the
behaviour of the wall is well predicted by the classical methods. Instead, when
increasing the anchor force by the double amount, the passive pressure may not be fully
mobilised and the failure mechanism differs form the classical predictions. In addition,
the anchor effect involves significant changes in the global stress state. Passive pressure
appears between the grout body and the wall, which creates a confined zone around the
anchor rod. This effect is not taken into account by the classical methods.
SAFETY FACTORS
tan
tan
1.1
SF14
SF13
SF12
tan
tan
SF15
tan
tan
1.1
tan
tan
1.1
SF16a
SF16b
tan 100 tan 200
tan
100 tan 200
1.1
1.3
1.1
Table 11. Effect of increasing the reference embedment depth by 20% for each method in
anchored walls.h=10m, d=2,5m
SAFETY FACTORS
SF17b
SF17a
tan
tan 100
1.3
tan
tan 200
1.4
SF18a
tan
tan 100
1.4
SF19a
SF18b
tan
tan 200
SF19b
tan
1.5
tan 100
1.5
SF20a
tan
tan 200
1.7
SF20b
tan
tan 100
1.6
tan
tan 200
1.7
Table 12. Safety provided by the classical methods in anchored walls. h=10m, d=2,5m
87
Conclusions
An evaluation of the safety provided by the classical methods on the design of sheet pile
walls has been carried out. The analytical results of six cases, comprising cantilever and
anchored walls, have been compared with the Finite Element model Plaxis.
The simplification of reality done by the classical methods allows us to generally
understand the behaviour of the whole body wall-soil. Nonetheless, the hypotheses in
which these are based on are found to be conservative. For instance, the hypothesis of
considering that both active and passive pressure are fully mobilised over the whole
length of the wall provides a lower embedment depth for the same soil friction angle.
Additionally, there are other effects that the classical methods ignore. Classical methods
may underestimate the passive pressure, as they do not take into account the
confinement in depth due to the three-dimensional stress state. Classical methods for
anchored walls neither take into account the confinement around the free length of the
anchor rod that increases the passive pressure.
The Finite Element numerical model requires knowing additional parameters, whose
influence needs to be known. Its interpretation requires understanding of the studied
phenomenon.
In cantilever walls, the real safety factor of each method comprises a range between 1.2
and 1.8, depending on the method. The most conservative is found to be the gradual
method with passive earth pressure.
In anchored walls, the real safety factor provided by the classical methods comprises a
range between 1.3 and 1.7 depending on the method and the anchor force.
For both cantilever and anchored walls, the traditional increase of 20% on the reference
embedment depth does not lead into the same amount on the safety factor. For each
method, the embedment increase of 20% provides a safety factor about 1.15 for
cantilever walls, and slightly less for anchored walls.
The active earth pressure is fully mobilised on the back of the wall above the dredge
level, even with soil friction angles far from the limit state. In contrast, the earth
pressure distribution on the back of the wall below the dredge level resembles the at-rest
state, with soil friction angles far from the limit state or with shallow depth of
excavation. The passive earth pressure is fully mobilised immediately below the dredge
level. However, its range and value increases with retained height, and exceeding the
theoretical value, due to the confinement caused by a three-dimensional stress state.
The failure wedge obtained by the classical methods and the Finite Element analysis are
not identical, even so they maintain a close similarity.
Even though the anchor force is increased, the overall stability is barely increased. An
anchor force increase of 200% does not lead into a double safety factor. For a
theoretical anchor force value, the classical methods predict the behaviour of the wall
fairly well. On the other hand, when the anchor force is increased, the stress state of the
soil is altered and the failure mechanism differs from the classical methods prediction.
Regardless of the anchor prestressing force, the final anchor load appears to trend to an
equilibrium load.
88
After comparing numerical results with the classical methods in anchored walls, the
failure mechanism given by the Finite Element analysis suggests an intermediate failure
state, where the sheet pile wall describes a translation movement on the deep zone, and
at the same time a rotation movement around the anchor point.
Small variations of the cohesion lead to a significant rise of the safety factor.
Consequently, ignoring this parameter provides reliability, but considering it wrongly
may have severe consequences.
References
36. BARDEN, L. (1974) Sheet pile wall design based on Rowes method. Part III of
A comparison of quay wall design methods. CIRIA Technical Note 54.
37. BICA, A. V. D. & CLAYTON, C. R. I. (1989) Limit equilibrium design methods
for free embedded cantilever walls in granular materials. ICE Proceedings, Part 1,
86, 879-898.
38. BLUM, H. (1931) Einspannungsverhltnisse bei Bohlwerken. W. Ernst & Sohn.
Berlin.
39. BOWLES, J. E. (1988) Foundation analysis and design. McGraw-Hill, 4th
Edition. New York.
40. BRANSBY, P. L. & MILLIGAN, G. W. E. (1975) Soil deformations near
cantilever sheet pile walls. Gotechnique 25, No. 2, 175-195.
41. CHERUBINI, C., GARRASI, A., PETROLLA, C. (1992) The reliability of an
anchored sheet pile wall embedded in a cohesionless soil. Canadian Geotechnical
Journal, 29, 426-435.
42. CHUGH, A. K. & LABUZ, J. F. (2011) Numerical simulation of an instrumented
cantilever wall. Canadian Geotechnical Journal, 48, 1303-1313.
43. DAY, R. A. & POTTS, D. M. (1993) Modelling sheet pile retaining walls.
Computers and Geotechnics, 15, 125- 143.
44. DAY, R. A. (1999) Net pressure analysis of cantilever sheet pile walls.
45. DAY, R. A. (2001) Earth pressure on cantilver walls at design retained heights.
ICE Proceedings Geotechnical Engineering, 149, Issue 3, 167-176.
46. FOURIE, A. B. & POTTS, D. M. (1989) Comparison of a finite element and
limiting equilibrium analyses for an embedded cantilever retaining wall.
47. HANNAH, T. H. (1987) Ground anchorages: anchor testing - measuring absolute
movement. ICE Proceedings, 82, Part 1, 639-644.
48. KING, G. J. W. (1995) Analysis of cantilever sheet-pile walls in cohesionless
soil. Journal of Geotechnical Engineering, Vol. 121, No.9.
49. MA, J., BERGGREN, B., BENGTSSON, P., STILLE, H., HINTZE, S. (2009)
Behaviour of anchored walls in soils overlying rock in Stockholm. International
Journal of Geoengineering Case Histories Vol. 2, Issue 1, 1-23.
50. MILLIGAN, G. W. E. (1983) Soil deformations near anchored sheet-pile walls.
51. OSMAN, A. S. & BOLTON, M. D. (2004) A new design method for retaining
walls in clay. Canadian Geotechnical Journal, 41, 451-466.
89
52. PADFIELD, C. J. &MAIR, R. J. (1984) Design of retaining walls embedded in

stiff clay. CIRIA Report 104.
53. PLANT, G. W. (1972) Anchor inclination - its effect on the performance of a
laboratory scale tied-back retaining wall. ICE Proceedings, Part 2.
54. Plaxis 8.2. (www.plaxis.nl)
55. POTTS, D. M. & FOURIE, A. B. (1984) The behaviour of a propped retaining
wall: results of a numerical experiment. Gotechnique 34, No.3, 383-404.
56. POWRIE, W. (1996) Limit equilibrium analysis of embedded retaining walls.
57. RICHARDS, D. J. & POWRIE, W. (1994) Finite element analysis of construction
sequences for propped retaining walls. ICE Proceedings, 107, 207-216.
58. ROWE, P. W. (1952) Anchored sheet-pile walls. ICE Proceedings. Part I, Vol. I.
59. ROWE, P. W. (1955) A theoretical and experimental analysis of sheet-pile walls.
ICE Proceedings. Part I, Vol. I.
60. ROWE, P. W. (1955) Sheet-pile walls encastr at the anchorage. ICE
Proceedings. Part I, Vol. I.
61. ROWE, P. W. (1956) Sheet-pile walls at failure. ICE Proceedings.
62. ROWE, P. W. (1957) Sheet-pile walls subject to line resistance above the
anchorage. ICE Proceedings.
63. ROWE, P. W. (1957) Sheet-pile walls in clay. ICE Proceedings. Part I, Vol. I.
64. SCHRIVER, A. B. &VALSANGKAR, A. J. (1996) Anchor rod forces and
maximum bending moments in sheet pile walls using the factored strength
approach. Canadian Geotechnical Journal, 33, 815-821.
65. SIMPSON, B. (2000) Partial factors: where to apply them? International
Workshop on Limit State Design in Geotechnical Engineering. Melbourne,
Australia.
66. KRABL, S. (2006) Interactional approach of cantilever pile walls analysis.
ActaGeotechnicaSlovenica, 2006/1.
67. USS (1975) Steel Sheet Piling Design Manual. United States Steel.
68. VALSANGKAR, A. J. &SCHRIVER, A. B. (1991) Partial and total factors of
safety in anchored sheet pile design. Canadian Geotechnical Journal, 28, 812-817.
69. YOUSSEF, Y. G. (2003) Berms for stabilizing earth retaining structures. Faculty
of Engineering of Cairo University. Fayoum, Egypt.
90
ANEEX II: BIBLIOGRAPHIC REVIEW
91
A summary of the most relevant literature is presented below.

Author
Theory
Engels
Design
method
Engels
Krey
Krey
Coulomb
Jacoby
Krey
Rankine
Lee
Krey
US Steel
Krey
Blum
Blum
Schmidt
Blum
Verdeyen
and Roisin
Blum
Rankine
Factor
of safety
Allowable
stress
1-3 on lower
end
1.2
embedment
Rankine
1.43
corrected for embedment
friction
Caquot and 1.2-1.4
on
Kerisel
embedment
Rankine
1.0
corrected for
friction
1.0
Costet and Blum

Sanglerat
CP-2
Blum
Rankine
active,
2
Rankine
passive
Rankine
active,
2
Rankine
passive
Caquot and
Kerisel
Packshaw
Padfield and Blum

Mair-Fp
Padfield and Blum
Mair-Fs
Padfield and Blum
Mair-Fr
Lechner
Caquot and
Kerisel
Caquot and
Kerisel
Caquot and
Kerisel
Coulomb
1.2-1.5 on Kp
Clayton and Blum

Milititsky
1.5 on Kp
Mayniel
active,
Caquot and
Kerisel
passive
Coulomb
1.5 on safe d,
And 1 for
Mmax
Caquot and 3 on tan and
Kerisel
tan for Kp
Brinch
1.2 to 1.3 on
Hansen
tan
Rowe
Rowe
Hong Kong Rowe

GCO
Danish
Brinch
Geotechnical Hansen
Institute
1.1-1.5
>25
Depth increase
Not necessary
Not necessary
2/3 to
0
passive
lower end
Not necessary
O active, 17 Not necessary
to /2 passive
17 for sand Not necessary
and steel
Approximated
Up to 0.4
by
wall
friction
correction
Approximated
2/3
by 1.2d
for -
2 on Kp
2 on Kp
1.2 on tan
1.5
1.0
92
Wall friction
2/3 active,
0.7 passive
15
active,
7.5 passive
2/3 active,
/2 passive
2/3 active,
/2 passive
2/3 active,
/2 passive
2/3
Approximated
by 1.2d
Approximated
by 1.2d
-
Approximated
by 1.2d
Approximated
by 1.2d
Approximated
by 1.2d
Approximated
by 1.2d
/2
active, Approximated
by 1.2d
2/3 passive
O for safe d, No
2/3 for Mmax
72
No
,
but Not necessary
mobilized
value might
Canadian
Blum
Geotechnical
Society
Rankine
1.0
active (see
ref.44
for
Kp)
be smaller
O active, /2- No
2/3 passive
Design methods for free embedded cantilever walls based on limit equilibrium
analyses
93
ANNEX III: STUDY OF THE K0 INFLUENCE
94
Cantilever sheet pile wall with 5 m of retained height and 5 m of embedment depth (h =
5 m, d = 5 m)
Results from software Plaxis
K0=0,3
-3
()
Stages
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
45
6,39
61,92
0,9995
ok
0,0005
67
40
14,62
86,73
ok
104
35
20,87
115,05
ok
130
30
32,74
162,09
0,9997
ok
0,0003
88
27
46,04
200,13
ok
93
26
52,39
218,49
0,9999
ok
1E-04
148
25
60,04
236,91
0,9995
ok
0,0005
165
Steps
24
73,76
257,86
0,9995
ok
0,0005
238
23
101,19
278,12
0,9967
col
0,0033
271
22
108,49
278,92
0,9742
col
0,0258
282
21
103,47
273,06
0,9444
col
0,0556
225
26
43,25
201,18
0,9329
col
0,0671
337
25
56,54
235,64
0,9994
ok
0,0006
422
24
61,49
243,4
0,9415
col
0,0585
489
23
49,29
219,06
0,697
col
368
22
76,1
259,74
0,7933
col
0,2067
506
()
Stages
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
Steps
45
6,81
63,01
0,9998
ok
0,0002
91
40
9,11
80,84
0,9991
ok
0,0009
59
35
13,2
110,5
ok
53
30
19,97
154,9
ok
54
27
28,4
200,93
ok
55
26
31,91
213,7
0,9994
ok
0,0006
60
25
36,79
234,8
0,9994
ok
0,0006
62
24
43,77
257,09
ok
67
23
56,39
284,43
ok
62
22
126,21
310,17
0,9994
ok
0,0006
74
21,5
96,13
300,17
0,986
col
0,014
87
21
98,86
294,96
0,9741
col
0,0259
79
20
214,43
288,27
0,951
col
0,049
76
25
36,84
235,45
ok
116
24
43,45
256,04
0,9998
ok
0,0002
144
23
54,16
280,47
0,9995
ok
0,0005
202
22
73,54
300,04
0,9634
col
0,0366
267
21
69,3
292,97
0,8151
col
0,1849
169
Staged Excavation
K0=1
-3
Staged Excavation
95
K0=2
-3
()
Stages
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
45
7,21
44,63
0,8557
col
0,1443
34
40
19,58
88,13
0,9991
ok
0,0009
107
35
25
108,8
0,9999
ok
1E-04
64
30
34,56
153,31
0,9995
ok
0,0005
69
27
45,51
199,81
ok
59
26
25
50,2
212,7
0,9999
ok
1E-04
56
57
Steps
57,17
237,08
ok
24
64,87
256,24
0,9996
ok
0,0004
55
23
79,96
283,78
ok
60
22
161,78
309,33
0,9993
ok
0,0007
68
21
266,18
294,42
0,9821
col
0,0179
85
25
56,09
234,95
0,9995
ok
0,0005
245
24
65,21
256,64
0,9997
ok
0,0003
291
23
77,89
280,92
0,9993
ok
0,0007
322
22
99,4
296,16
0,9481
col
0,0519
432
21
90,01
285,03
0,7825
col
0,2175
340
()
Stages
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
Steps
45
2,439
-29,74
0,1504
col
0,8496
13
40
1,906
-33,54
-0,0361
col
1,0361
35
3,055
-25,34
0,091
col
0,909
30
49,42
156,28
ok
92
27
54,61
197,74
0,9998
ok
0,0002
77
26
58,25
210,95
0,9994
ok
0,0006
79
25
63,58
235,55
0,9994
ok
0,0006
75
Staged Excavation
K0=5
-3
24
70,42
256,9
ok
77
23
83,62
281,38
0,9997
ok
0,0003
75
22
139,92
305,79
0,9987
col
0,0013
101
21
149,04
289,66
0,9835
col
0,0165
81
Staged Excavation
26
59,55
217,41
0,9993
ok
0,0007
273
25
63,46
235,72
ok
278
24
70,81
256,17
0,9992
ok
0,0008
310
23
83,54
280,85
0,9998
ok
0,0002
342
22
100,45
293,64
0,9412
col
0,0588
427
96
Case form A. Cuadrardo (2010)

Stages
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
45
3,76
56,67
0,9997
ok
0,0003
60
40
6,51
78,55
ok
64
33
14,81
127,15
ok
73
31
19,18
143,94
0,9994
ok
0,0006
71
30,5
30
150,77
156,65
1
0,9995
ok
20,99
22,76
ok
0
0,0005
105
81
29,5
24,74
164,41
0,9997
ok
0,0003
67
29
27,07
174,09
ok
92
28,5
29,34
181,6
0,9999
ok
1E-04
87
Steps
28
31,67
187,95
0,9991
ok
0,0009
86
27,5
34,46
195,21
0,9993
ok
0,0007
105
27
37,11
199,97
0,9997
ok
0,0003
140
201,03
0,9995
91
26,9
K0=0,5
-3
()
37,79
ok
0,0005
26,8
38,35
202,25
0,9993
ok
0,0007
100
26,7
38,94
203,13
0,9995
ok
0,0005
136
26,6
39,9
206,11
0,9996
116
40,59
208
0,9993
ok
ok
0,0004
26,5
0,0007
140
26
44,17
217,22
0,9995
ok
0,0005
134
25,5
48,36
226,8
0,9992
ok
0,0008
148
25
53,38
235,8
0,9992
ok
0,0008
153
24
70,17
256,57
0,9992
ok
0,0008
258
23
1
1
83,48
265,09
0,9867
col
0,0133
259
94,92
269,52
0,968
col
0,032
283
27
35,35
197,71
ok
277
26
25
42,05
214,08
ok
329
51,7
234,98
0,9993
ok
0,0007
425
24
67,2
255,25
0,9994
ok
0,0006
552
23
74,86
260,41
0,9191
col
0,0809
555
22
70,54
255,29
0,7687
col
0,2313
477
()
Stage
s
Ux (10 m)
Mmax (kNm)
Mstage
State
1-Mstage
Steps
45
5,09
ok
80
40
7,2
ok
79
35
10,96
ok
66
30
18,07
ok
56
27
26,25
ok
81
26
30,96
ok
64
25
36,73
ok
70
24
46,1
0,9996
ok
0,0004
74
23
64,13
ok
83
22
95,46
0,9907
col
0,0093
194
21
79,15
0,9612
col
0,0388
134
22
Staged Excavation
K0=0,75
-3
Staged Excavation
25
ok
24
ok
23
ok
22
col
21
col
97
K0=1,5
-3
()
Stages
Ux (10 m)
Mstage
State
1-Mstage
Steps
45
9,53
0,9775
col
0,0225
111
40
14,15
0,9994
ok
0,0006
100
35
18,98
ok
59
30
26,8
0,999
ok
0,001
62
27
36,78
ok
51
26
40,98
ok
51
25
47,13
ok
52
24
54,24
0,9991
ok
0,0009
57
23
69,98
0,9998
ok
0,0002
59
22
150,48
0,9996
ok
0,0004
66
21
369,76
0,9808
col
0,0192
85
Mmax (kNm)
Staged Excavation
25
ok
24
ok
23
ok
22
col
21
col
130
120
110
100
Ux attop(mm)
90
80
Ko=0,3
70
Ko=0,5
Ko=0,75
60
Ko=1
50
Ko=1,5
40
Ko=2
30
Agustn
20
10
0
15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45
()
Evolution of the movement of the top of the wall when soil friction angle is decreased for
the geometry h = 5 m, d = 5 m
98
Results for the same geometry h = 5 m d = 5 m. Relation between K0 and Ux for a given friction
angle.
-3
K0
0,25
0,5
0,7
0,75
0,8
1
1,25
1,5
1,75
()
30
30
30
30
30
30
30
30
30
Stages
1
1
1
1
1
1
Ux (10 m)
32,71
21,71
17,79
18,07
18,18
19,97
Mmax (kNm)
161,96
156,15
154,68
156,01
155,19
154,9
Mstage
0,9999
0,9998
1
1
1
1
1
1
1
23,41
26,8
30,85
155,8
153,66
154,4
30
34,56
K0
0,405
0,5
0,7
0,75
0,85
1
1,25
1,5
1,75
()
25
25
25
25
25
25
25
25
25
Stages
1
1
1
1
1
1
1-Mstage
1E-04
0,0002
0
0
0
0
0
0,001
0
Steps
119
72
64
56
68
54
1
0,999
1
State
ok
ok
ok
col
ok
ok
ok
ok
ok
56
62
64
Ux/max
31,7%
21,1%
17,3%
17,5%
17,6%
19,4%
22,7%
26,0%
29,9%
153,31
0,9995
ok
0,0005
69
33,5%
Ux (10 m)
60,23
51,88
38,33
36,73
35,92
36,79
Mmax (kNm)
237,59
235,65
41,84
47,13
51,95
236,59
136,46
236,2
1
1
1
1-Mstage
0,0006
0,0009
0
0
0
0,0006
0
0
0
Steps
174
176
85
70
82
62
1
1
1
State
ok
ok
ok
ok
ok
ok
ok
ok
ok
Ux/max
58,4%
50,3%
234,8
Mstage
0,9994
0,9991
1
1
1
0,9994
57
52
55
35,7%
40,6%
45,7%
50,4%
25
57,17
237,08
ok
57
55,4%
K0
0,25
0,5
0,75
1
1,25
1,5
1,75
()
35
35
35
35
35
35
35
Stages
1
1
1
1
Ux (10 m)
21,16
10,72
10,96
13,2
Mmax (kNm)
114,01
106,51
108,49
110,5
Mstage
0,9996
1
1
1
15,65
18,98
22,11
110,11
113,5
113,45
1
1
0,9997
1-Mstage
0,0004
0
0
0
0
0
0,0003
Steps
75
60
66
53
1
1
1
State
ok
ok
ok
ok
ok
ok
ok
75
59
53
Ux/max
20,5%
10,4%
10,6%
12,8%
15,2%
18,4%
21,4%
35
25
108,8
0,9999
ok
1E-04
64
24,2%
-3
-3
235,49
35,6%
Viability of cases depending on the initial stress state:

()
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
Ka
0,1716
0,1802
0,1891
0,1982
0,2077
0,2174
0,2275
0,2379
0,2486
0,2596
0,2710
0,2827
0,2948
0,3073
0,3201
0,3333
0,3470
Kp
5,8284
5,5500
5,2893
5,0447
4,8150
4,5989
4,3955
4,2037
4,0228
3,8518
3,6902
3,5371
3,3921
3,2546
3,1240
3,0000
2,8821
K0=0,3
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
NO
NO
NO
NO
K0=0,485
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
99
K0=1
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
K0=2
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
K0=5
OK
OK
OK
OK
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
Numerrical analysis of cantilever andd anchored sheeet pile walls at failure

f
and com
mparison with cllassical method
ds
28
27
26
25
24
23
22
21
20
19
18
17
16
15
0,3610
0,3755
0,3905
0,4059
0,4217
0,4381
0,4550
0,4724
0,4903
0,5088
0,5279
0,5475
0,5678
0,5888
2,7698
2,6629
2,5611
2,4639
2,3712
2,2826
2,1980
2,1171
2,0396
1,9655
1,8944
1,8263
1,7610
1,6984
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
OK
OK
OK
OK
OK
OK
OK
OK
NO
NO
NO
NO
NO
NO
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
NO
NO
NO
NO
NO
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
NO
N
14
0,6104
1,6383
NO
NO
OK
NO
NO
N
Horizontalmovementattop
(mm)
(mm)
70
60
50
40
30
20
10
0
0
0,25
0,5
5
0,75
1,25
1,5
5
1,75
2,25
2
K0
Relatiion between
n K0 and horrizontal movvement
Canttilever sheett pile wall with

w 3 m off retained heeight and 5,44 m of embbedment dep
pth (h
= 3 m,
m d = 5,4 m).
m
K0=0
0,25
Stages
S
1
1
1
1
1
1
1
1
1
1
1
1
() Ux (10-3
m) Mmax (kNm) Mstag
ge Estado 1-Mstage Steps
35
2,78
8
25,01
1
ok
0
9
30
4,23
3
36,65
5
1
ok
0
28
25
7,98
8
53,74
4
0,9999
9
ok
1E-04
72
23
9,81
1
63,3
3
1
ok
0
94
21
11,9
95
0,999
0,001
93
20
13,3
38
83,39
9
0,9994
4
ok
0,0006
107
19
15,2
21
92,16
6
0,9996
6
ok
0,0004
143
18
17,7
72
103,3
32
1
ok
0
146
17
21,3
35
116,8
82
0,9997
7
ok
0,0003
141
16
28
132,4
49
1
ok
0
56
15
45,14
150,4
44
0,9991
1
0,0009
184
14
166,9
95
159,1
15
0,9784
4
col
0,0216
79
Staged
d Excavation
3
15
3
14
45,14
166,9
95
150,4
44
159,1
15
100
0,9991
1
0,9784
4
col
0,0009
0,0216
225
125
K0=0,5
Stages
1
1
1
1
1
1
1
1
1
1
1
1
1
() Ux (10-3m) Mmax (kNm) Mstage Estado 1-Mstage Steps

40
18,53
0,9997
ok
0,0003
15
35
23,14
0,9991
ok
0,0009
17
30
1,565
33,6
1
ok
0
24
25
4,38
51,67
0,9997
ok
0,0003
39
23
7,11
62,44
1
ok
0
77
21
10,99
75,65
1
ok
0
94
20
12,94
83,6
0,9997
ok
0,0003
110
19
15,08
92,45
1
ok
0
121
18
17,5
103,32
0,9997
ok
0,0003
128
17
21,11
116,95
0,9995
ok
0,0005
146
16
27,41
132,16
0,9994
ok
0,0006
163
15
45,41
150,8
1
ok
0
66
14
4180
160,58
0,9811
col
0,0189
304
Staged Excavation
3
15
3
14
K0=0,75
Stages
1
1
1
1
1
1
1
1
1
1
1
1
45,41
4180
150,8
160,58
1
0,9811
ok
col
0
0,0189
103
337
-3
() Ux (10 m) Mmax (kNm) Mstage Estado 1-Mstage Steps
35
1,141
27,03
1
ok
0
17
30
1,79
34,66
1
ok
0
25
25
3,13
50,1
1
ok
0
33
23
4,14
60,28
0,9998
ok
0,0002
33
21
5,68
73,85
1
ok
0
39
20
6,761
82,07
1
ok
0
45
19
8,26
91,49
1
ok
0
46
18
10,32
103,23
1
ok
0
45
17
13,29
116,03
0,9998
ok
0,0002
51
16
18,52
131,82
1
ok
0
53
15
31,51
150,39
0,9996
ok
0,0004
60
14
496,72
160,3
0,9808
col
0,0192
78
Staged Excavation
3
15
3
14
31,51
496,72
150,39
160,3
101
0,9996
0,9808
ok
col
0,0004
0,0192
93
117
-3
() Ux (10 m) Mmax (kNm) Mstage Estado 1-Mstage Steps
35
0,679
23,17
1
ok
0
16
30
1,62
33,71
1
ok
0
24
25
4,9
51,97
1
ok
0
51
20
13,26
83,5
1
ok
0
127
19
15,14
92,65
1
ok
0
52
18
17,46
103,15
0,9993
ok
0,0007
125
17
21,01
116,4
0,9996
ok
0,0004
147
16
27,31
131,78
0,9993
ok
0,0007
155
15
45,36
150,74
1
ok
0
65
14
151,54
158,88
0,9786
col
0,0214
76
Stages
1
1
1
1
1
1
1
1
1
1
K0=0,485
Staged Excavation
3
3
3
3
16
15
14
13
27,31
45,36
151,54
351,99
131,78
150,74
158,88
151,49
0,9993
1
0,9786
0,9282
ok
ok
col
col
0,0007
0
0,0214
0,0718
189
94
108
116
35
Uxattop(mm)
30
25
20
15
Ko=0,5
10
Ko=0,75
5
0
0
10
15
20
25
30
35
40
45
()
Evolution of the horizontal movement as soil friction angle is decreased for the
geometry h = 3 m, d = 5.4 m
102
ANNEX IV: MULTIPLE-ANCHORED SHEET PILE WALLS
103
Results form Plaxis for the three-level-anchored sheet pile wall. Retained height h =
12m. Embedment depth d = 3 m.
Tables below show all the cases considered and the stages followed in Plaxis. In the
right column are the results.
CASE1"Ah12d3simultaneoETAPASKo136"
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
530,44
83,83
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Prestress125kN/mupperanchor
10
6,5
STAGE4
Excavation1m
STAGE5
Excavation1m
Displ.Hzt.TOP
Displ.Hzt.TOE
STAGE6
Excavation1m
68
16,86
STAGE7
Excavation1m
STAGE8
Excavation1m
Mupperanchor
Mloweranchor
STAGE9
Prestress125kN/mloweranchor
44
145,55
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
Excavation1m
STAGE13
Excavation1m
STAGE14
Excavation1m
CASE2"Ah12d3CASE2Ko136"
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
533,25
89,44
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
STAGE5
Excavation1m
STAGE6
Excavation1m
STAGE7
Excavation1m
STAGE8
Excavation1m
STAGE9
Excavation1m
STAGE10
STAGE11
Excavation1m
STAGE12
10
Displ.Hzt.TOP
Displ.Hzt.TOE
75,78
16,66
Mupperanchor
Mloweranchor
47,32
160,19
Excavation1m
STAGE13
Excavation1m
STAGE14
Excavation1m
104
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
525,96
101,96
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
STAGE5
STAGE6
Excavation1m
STAGE7
Excavation1m
STAGE8
Excavation1m
10
Displ.Hzt.TOP
Displ.Hzt.TOE
93,28
16,16
Mupperanchor
Mloweranchor
STAGE9
Excavation1m
STAGE10
Excavation1m
STAGE11
STAGE12
Excavation1m
STAGE13
Excavation1m
STAGE14
Excavation1m
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
546,64
133,51
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
STAGE5
Excavation1m
STAGE6
STAGE7
Excavation1m
STAGE8
Excavation1m
46,73
187,39
10
3,125
Displ.Hzt.TOP
Displ.Hzt.TOE
130,7
15,16
Mupperanchor
Mloweranchor
STAGE9
Excavation1m
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
STAGE13
Excavation1m
STAGE14
Excavation1m
105
48,58
239,05
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
Displ.Hzt.TOE
STAGE5
Excavation1m
Displ.Hzt.TOP
STAGE6
STAGE7
Excavation1m+P1
STAGE8
Excavation1m+P1
COLLAPSE
Mupperanchor
Mloweranchor
STAGE9
Excavation1m+P1
STAGE10
Excavation1m+P1
STAGE11
Excavation1m+P1
STAGE12
STAGE13
Excavation1m+P1+P2
STAGE14
Excavation1m+P1+P2
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
499,24
183,47
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
STAGE5
Excavation1m
STAGE6
Excavation1m
STAGE7
STAGE8
Excavation1m
9,875
Displ.Hzt.TOP
Displ.Hzt.TOE
183,47
14,47
Mupperanchor
Mloweranchor
STAGE9
Excavation1m
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
STAGE13
Excavation1m
STAGE14
Excavation1m
106
48,99
176,16
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
670,66
305,25
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
STAGE5
Excavation1m
STAGE6
Excavation1m
STAGE7
Excavation1m
STAGE8
9,375
Displ.Hzt.TOP
Displ.Hzt.TOE
305,25
11,48
Mupperanchor
Mloweranchor
STAGE9
Excavation1m
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
Excavation1m
STAGE13
STAGE14
Excavation1m
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
Displ.Hzt.TOP
Displ.Hzt.TOE
STAGE5
Excavation1m
STAGE6
Excavation1m
STAGE7
Excavation1m
STAGE8
Excavation1m
46,78
463,95
COLLAPSE
Mupperanchor
Mloweranchor
STAGE9
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
Excavation1m
STAGE13
Excavation1m
STAGE14
107
CASE6.1"Ah12d3CASE6.1Ko136"
MMAX
Displ.Hzt.MAX
STAGE1
Excavation1m
STAGE2
Excavation1m
location(fromtop)
location(fromtop)
STAGE3
Excavation1m
STAGE4
Excavation1m
Displ.Hzt.TOE
STAGE5
Excavation1m
Displ.Hzt.TOP
STAGE6
Excavation1m
STAGE7
STAGE8
Excavation1m
COLLAPSE
Mupperanchor
Mloweranchor
STAGE9
Excavation1m
STAGE10
Excavation1m
STAGE11
Excavation1m
STAGE12
Excavation1m
STAGE13
STAGE14
Excavation1m
Relative shear stresses around a three-level-anchored sheet pile wall. Dark and clear
shadings represent lower and higher values of stress respectively.
Relative shear stresses around a two-level-anchored sheet pile wall

108
Shear strains around a three-level-anchored sheet pile wall. Dark and clear shadings
represent lower and higher values of strain respectively. Maximum strain 7,04% scale
0% to 3,6%
Shear strains around a two-level-anchored sheet pile wall. Maximum strain 8,46% 0%
to 3,6%
109

Numerical Analysis of Cantilever and Anchored Sheet Pile Walls

Uploaded by

Copyright:

Available Formats

Numerical Analysis of Cantilever and Anchored Sheet Pile Walls

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Numerical Analysis of Cantilever and Anchored Sheet Pile Walls

Uploaded by

Copyright:

Available Formats

Terms and Conditions of Use:

this document downloaded from

All of the information, data and computer software

And much more...

NUMERICAL ANALYSIS OF CANTILEVER AND

ALEJO GONZLEZ TORRABADELLA

ALEJANDRO JOSA GARCA-TORNEL

Gracias a mi familia, a mis compaeros y a mis amigos. Gracias a mi madre por

1. INTRODUCTION AND OBJECTIVES .................................................................. 3

First sight on sheet pile walls ............................................................................. 5

Global mechanical behaviour ............................................................................ 6

Stress state .................................................................................................. 6

Earth pressures ................................................................................................... 8

Coulomb Theory ......................................................................................... 8

Rankine Theory .......................................................................................... 9

Classical design methods ................................................................................. 10

Cantilever walls ........................................................................................ 11

Single-anchored walls .............................................................................. 13

Multiple-anchored walls ........................................................................... 15

3. BIBLIOGRAPHIC REVIEW ................................................................................. 18

Free embedded cantilever sheet pile walls....................................................... 18

Anchored sheet pile walls ................................................................................ 29

4. NUMERICAL MODEL .......................................................................................... 34

Modelling sheet pile walls ............................................................................... 35

Loads and boundary conditions ................................................................ 38

Mesh and initial stresses generation ......................................................... 41

5. INFLUENCE OF THE INITIAL STRESS STATE ............................................... 44

Analytical methods .......................................................................................... 46

Results analysis ................................................................................................ 46

6. MULTIPLE-ANCHORED SHEET PILE WALLS ................................................ 49

Geometry and methodology ............................................................................. 51

Study of the influence of the construction procedure ............................... 51

Study of anchor load variation ................................................................. 52

Three-level-anchored sheet pile wall ........................................................ 53

Results analysis ................................................................................................ 54

1. INTRODUCTION AND OBJECTIVES

The process has consisted of the following chronological steps:

First sight on sheet pile walls

Retaining structures are commonly divided into two families:

Resist the thrust generated by the excavation.

Figure 2.1 Cantilever sheet pile wall

Global mechanical behaviour

The main feature of an embedded retaining structure is the contribution on stability of

The type of earth pressure distribution

2.2.1. Stress state

2.3.1. Coulomb Theory

Where the angles , and , as seen in Figure 2.5, represent:

: Inclination of the wall from the horizontal.

Figure 2.5 Coulomb wedge analysis

2.3.2. Rankine Theory

Classical design methods

2.4.1. Cantilever walls

Figure 2.10 Simplified method. The thrust below the rotation

Figure 2.11 Gradual method, shown in net pressures

2.4.2. Single-anchored walls

Fixed earth support method

Figure 2.14 Fixed earth support method

This method leads to a maximum reference embedment depth. It requires an additional

2.4.3. Multiple-anchored walls