Thesis Report Final FH Leferink

Delft University of Technology
MSc Thesis report
Research into the application of

undrained analysis with critical state
soil mechanics approach for the design
of a quay wall
Graduation Committee:
Author:
Prof. Dr. Ir. S.N. Jonkman Delft University of Technology
F.H. Leferink
Ir. J.H. van Dalen Delft University of Technology
Dr. Ir. J.G. de Gijt Delft University of Technology
Dr. Ir. T. Schweckendiek Delft University of Technology
Ir. H.E. Pacejka Municipality of Rotterdam
February 3, 2020
2
Abstract
The design and construction of quay walls are processes that exist for many centuries and
have become more complex and challenging in current engineering practice.
For the design of quay walls a number of guidelines and design codes have been devel-
oped over the years. These give the requirements that a quay wall structure should meet,
but do also provide some guidance in which steps to take in order to arrive at a proper
final design. The relevance of undrained soil behavior, described using critical state soil
mechanics, for the analysis of quay wall stability is yet unknown. The main objective
of this research is to investigate the possibilities to use the alternative design approach
for modelling soil behavior in the design processes of a quay wall. For this, three case
studies have been elaborated. The three case studies represents soil profiles consisting
1) predominantly sandy soils, 2) normally consolidated clay and 3) overconsolidated clay.
The differences between the analyses and outcomes of the conventional approach and the
new approach have been compared for each case study.
Based on the quantitative results of case study 1 and 3, the difference in outcome be-
tween the conventional and alternative design approaches is between 0 and 10% for both
displacements and sectional forces. The outcome of case study 2 is not in line with the
results of case 1 and 3.
Based on the results of the first case study with sandy soil profile, the alternative design
approach applied in this report is not a valid option for the design of a quay wall due to
the absence of undrained soil conditions. For a soil profile consisting clay, the magnitude
of preconsolidation of the soil plays an important role. For the alternative design ap-
proach, increasing values of pre-loading results in decreasing values of sectional forces and
displacements of the wall. This effect is stronger in comparison to the conventional design
approach. In further research the aim should be to increase the reliability of the alternative
design approach. This can be done by using in-situ measurements of the displacements of
the wall to validate if the model represents the reality accurately.
3
4
Contents
Abstract 3
1 Introduction 8
1.1 Problem description & research questions . . . . . . . . . . . . . . . . . . . 9
1.2 Method and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Theoretical background 10
2.1 Design of quay walls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Soil mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.1 Stress–strain relations . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 Shear strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.3 Horizontal soil stresses . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.4 Undrained soil behavior . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Design methods for quay walls . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.1 Method of Blum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Spring-supported elastic beam model . . . . . . . . . . . . . . . . . . 17
2.3.3 FEM - Plaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Developments in dike safety assessment . . . . . . . . . . . . . . . . . . . . 20
2.4.1 Design process for a dike vs quay wall . . . . . . . . . . . . . . . . . 20
2.4.2 New design approach for dikes based on critical state soil mechanics 21
2.5 Material models for describing soil behavior . . . . . . . . . . . . . . . . . . 23
2.5.1 Mohr-Coulomb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.2 Hardening Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.3 Soft Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5.4 NGI-ADP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5.5 Shansep NGI-ADP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3 Method 27
3.1 Framework case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1.1 General modelling choices . . . . . . . . . . . . . . . . . . . . . . . . 29
3.1.2 Specific modelling choices per case . . . . . . . . . . . . . . . . . . . 31
3.2 Framework for determining soil parameters . . . . . . . . . . . . . . . . . . 32
3.2.1 General soil parameters . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2.2 Shansep model parameters . . . . . . . . . . . . . . . . . . . . . . . 33
5
3.2.3 NGI-ADP model parameters . . . . . . . . . . . . . . . . . . . . . . 34
4 Case study 1 - Quay wall Amazonehaven 36

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2 Parameters / Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.1 Soil profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.3 Design verification calculation based on theory of Blum . . . . . . . . . . . 44
4.4 Designs based on conventional design approach . . . . . . . . . . . . . . . . 45
4.5 Design based on alternative design approach using cssm . . . . . . . . . . . 48
4.5.1 Design using D-Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.5.2 Design using Plaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.6.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.6.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5 Case study 2 - Amazonehaven modified with normally consolidated clay

layer 55
5.1 Parameters / Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . 56
5.1.1 Type of soil response . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.1.2 Soil profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1.3 Different scenarios regarding the timing of construction and loading
of the wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Design based on alternative design approach using undrained analysis com-
bined with cssm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.4.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.4.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6 Case study 3 - Quay wall situated in overconsolidated clay layer 67

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.2 Parameters and Boundary conditions . . . . . . . . . . . . . . . . . . . . . . 68
6.2.1 Type of soil response . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.2.2 Soil profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.3.1 Preliminary design based on theory of Blum . . . . . . . . . . . . . . 71
6.4 Design based on alternative design approach using undrained analysis com-
bined with cssm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.5 Sensitivity analysis for historic loads . . . . . . . . . . . . . . . . . . . . . . 76
6.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.6.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.6.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7 Discussion 82
8 Conclusion 84
6
List of Figures 86
List of Tables 88
A Info about current practise CSSM in dikes and embankments 92

A.1 Approach for determining the critical state friction angle . . . . . . . . . . . 92
B Finite Element Method 95
C Additional information case study 1: Amazonehaven 97

C.1 D-Sheet calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
D Additional information case study 2: soft soil consisting of clay and peat 98
D.1 Soil investigation data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
E Additional information case study 3: Boulder clay 105

E.1 Soil investigation data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7
CHAPTER 1
Introduction
The design, construction and use of quay walls are processes that exist already for many
centuries. In time, more and more knowledge and experience about quay walls has de-
veloped, like the strength, safety, reliability and structural behavior of the wall and the
surrounding soil, to name a few. Research into these topics is still being conducted up to
this day. One of these research areas is the soil-structure interaction between quay walls
and the surrounding soil, and how soil behavior affects this interaction. Sufficient knowl-
edge about the modelling of this interaction will contribute to a safe and cost-efficient
design.
In 2017, an alternative approach has been introduced in the Netherlands for the assessment
of slope stability of dikes. The prevailing approach based on the theory of Mohr-Coulomb
(MC) has been replaced by a new method considering an undrained soil response and using
critical state soil mechanics theory (CSSM). Research of several case studies has shown
that the application of CSSM in combination with undrained analysis of the soil gives a
more accurate description of the behaviour of soft organic clay and peat compared to the
conventional approach (van Duinen and van Hemert, 2013). Also, using effective stress
shear strength parameters based on triaxial tests often produces high stability factors,
which did not always comply with reality (van Duinen, 2013).
Research into this subject is still ongoing, for instance in the Dutch research project
Projectoverstijgende Verkenning Macrostabiliteit (POVM). One of the research areas is the
development and improvement of design tools for assessing the slope stability of the dikes,
also taking into account possible structural elements in the dike. The structural elements,
for instance sheet pile walls, contribute to the resistance of several failure mechanisms,
including slope stability of the dike. To model this soil-structure interaction, often a
Finite Element Method is used.
With respect to the slope stability of the dike, the sheet pile wall has two objectives: 1)
decreasing the amount of seepage through the dike and 2) soil retention in case of shearing
of the soil. In the last case, the sheet pile wall intersects the potential sliding plane of
the soil, and the bending stiffness of the structural element will prevent the soil body
from sliding. The second objective mentioned above is one of the primary functions of
a quay wall. Despite of differences between a dike body and a quay wall in for instance
geometry and load cases, there are also similarities. For example, in both cases a difference
in horizontal pressure across the structure is present which greatly affects the functioning
8
of the structure. To gain further insight in the slope stability of quay walls, different
constitutive models for describing soil behavior will be investigated.
1.1 Problem description & research questions

For the design of quay walls a number of guidelines and design codes have been developed
over the years. These give the requirements that a quay wall structure should meet, but
do also provide some guidance in which steps to take in order to arrive at a proper final
design. The relevance of undrained soil behavior, described using CSSM, for the analysis
of quay wall stability is yet unknown.
The main objective of this research is to investigate the possibilities to use the alternative
design approach for modelling soil behavior in the design processes of a quay wall. This
objective can be summarized in the following research question:
Can critical state soil mechanics in combination with undrained analysis of the soil be
applied for the design of quay walls, and what are the effects for the prediction of strength
and stability?
In order to be able to answer the research question, the following sub questions have
to be addressed:
1. For which situations / conditions is the new assessment approach for dikes developed,
and do these conditions also occur in case of retaining structures?
2. Which failure mechanisms of quay walls can be analyzed using the new approach?
3. How do the results of the new approach compare to the conventional design ap-
proach?
1.2 Method and scope

To identify the possibilities for application of this approach for quay walls, three case
studies have been elaborated. For each case study, two designs are elaborated. The first
design is based on the conventional design approach, following the guidelines and design
rules CUR166 (2008) and CUR211 (2014). For the second design the new approach is
applied, following the regulations and guidelines prescribed in WBI2017. The scope of
this research is the soil characteristics and loading situations for which the new approach
could be applied. The three case studies represents soil profiles consisting 1) predomi-
nantly sandy soils, 2) normally consolidated clay and 3) overconsolidated clay.
Using different material models in Plaxis, quantities such as the strength and stability
of a soil body, the displacements of soil bodies or structures and structural forces can
be calculated. Also, this program is able to take into account the effects of pore water
pressures and therefore suited to model undrained soil behavior.
The differences between the analyses and outcomes of the conventional approach and the
new approach will be compared for each case study.
9
CHAPTER 2
Theoretical background
In this chapter, general background information regarding the design of quay walls is
presented. This entails the general characteristics and types of quay walls (section 2.1),
basic soil mechanics (section 2.2) and different methods to design and calculate a quay wall
structure (section 2.3). Besides this, the new design approach for dikes in the Netherlands
is briefly explained in section 2.4. Lastly, in section 2.5 some technical information about
the material models used in the FEM software plaxis is presented.
2.1 Design of quay walls

This section provides general background information about quay walls, i.e. functions
of quay walls, different types of quay wall structures and different failure mechanisms
that can occur. Most part of this section is based on the design guidelines CUR166 and
CUR211.
Functions of quay walls

Quay walls are structural elements that enable ships to moor and to exchange cargo and
passengers with the land. This can be inside an harbour, along a river or navigation
canal but also along canals inside city centres. The quay structure must meet all kind of
requirements based on functionality, user demands, soil conditions, water levels, size of
the vessels, magnitude of external loads due to mooring and transferring cargo. The main
functions of quay walls are (CUR211, 2014):
• Providing berthing facilities for ships
• Retaining the soil body located behind the quay
• Providing the bearing capacity to carry the loads imposed by transshipment of cargo,
presence of cranes, other equipment and freight storage facilities
• Act as flood defense for the area behind the quay during periods of high water
Types of quay walls

Given a set of specific functions and boundary conditions that have to be fulfilled, a choice
can be made for the design between different types of quay walls. Each type (or combi-
10
nation of types) has certain characteristics which makes it more suitable for particular
design conditions. The main types of quay walls are being described below.
Gravity walls
These structures ensure stability by their self weight, possibly added with the pressure of
the soil on top of the structure.
Sheet pile walls

Sheet pile walls ensure stability from the bending capacity of the vertical elements making
up the wall, which are fixated deep inside the soil. The elements are connected using inter-
locks, to ensure soil-tightness. The wall often consists of a combined wall of steel tubular
piles with sheet piles in between. The wall can also be anchored to create a support at
the top. This type is most common for quay walls in the Netherlands.
Structures with relieving platforms

In this type of quay wall a concrete superstructure is constructed on top of the combined
wall at the waterside, and additional support is created by applying concrete bearing piles
and anchorages. The structure will reduce the amount of soil behind the quay wall, and
will transfer the surcharge loads effectively to the load bearing soil layers, thereby signifi-
cantly reducing the horizontal soil pressures behind the wall.
Open berth quays (jetties)

In this type of structure, the soil is not retained by a vertical wall but by a slope beneath
the berthing platform, covered with revetment. The platform is supported by vertical or
inclined bearing piles, and possibly anchorages. This type has the advantage that the soil
pressures on the structure are much lower compared to the other types, but more space is
required.
Failure mechanisms of quay walls
Depending on the type of quay wall structure, multiple failure mechanisms can occur.
For each of these, the requirements for structural safety (ultimate limit state, ULS) or
usability (serviceability limit state, SLS) of the structure have fallen short. The most
relevant failure mechanisms of a quay wall structure are:
• Exceeding of the horizontal or vertical bearing resistance of the soil
• Exceeding of the resistance for shearing of the soil
• Exceeding of the strength of the wall due to (a combination of) bending moment,
normal forces and shear forces
• Exceeding of the strength of anchorage
• Exceeding of the strength of bearing piles
• Loss of global stability of the structure
• Exceeding of the allowable deformations of the wall
The design standards Eurocode provide a set of requirements for both ULS and SLS
to ensure a certain safety level for the structure. Design guidelines such as CUR166 and
CUR211 are based on the Eurocode and have at some points extended requirements. Also,
for most designs of quay walls specific requirements have been established in agreement
with the client, based on the desired functionality and usability.
11
2.2 Soil mechanics
2.2.1 Stress–strain relations
In general, a soil consist of a mixture of soil particles, water and air. The space between
the soil particles are called pores, and in case of a saturated soil the pores are filled with
water. The magnitude of pores of a soil is expressed in the variable of porosity n, and this
is an essential property of soils and soil behavior.
The deformations of soils are determined by the effective stresses, which are a measure for
the contact forces transmitted between the particles. If the ratio of shear force and normal
force exceeds a certain value (the friction coefficient of the particles), the particles will
start to slide over each other, which will lead to deformations. The deformations caused by
shearing of the soil are usually much larger than the deformations in compression. Also, in
compression the material becomes gradually stiffer, whereas in shear it becomes gradually
softer. This is the stress-dependency of stiffness, and is an essential property of granular
soils.
Deformations of soils are irreversible. After a full cycle of loading and unloading of a soil,
a permanent deformation is observed. The behavior in unloading and reloading, below
the maximum load sustained before, often seems practically elastic, with some additional
plastic deformation after each cycle. A soil is said to be overconsolidated if it has been
loaded before, by a certain preload. Up to the stresses of the preload, the soil will behave
relatively stiff. When the stresses exceed the stress of the preload, the soil will behave
relatively soft again.
A very characteristic property of relatively dense soils is the ability to increase in volume if
loaded in shear. This is called dilatancy. If dilatancy occurs in a saturated soil, water must
be attracted to fill the additional pore space. If this process cannot occur fast enough, an
under pressure in the pores can occur, resulting in temporary higher effective stresses. The
reverse effect can occur in case of very loosely packed sand, and this is called contractancy.
At continuing deformations both dense and loose sand will tend towards a state of average
density, sometimes denoted as the critical density. At high stresses the critical density is
somewhat smaller than at small stress.
2.2.2 Shear strength

The shear strength of a soil is generally defined by Coulomb:
τ = c0 + σv0 · tan φ0 (2.1)
Combined with Mohr’s circle this results in the Mohr-Coulomb (MC) failure criteria. The
MC-failure criteria are used in several models for describing soil behavior. Equation 2.2
gives the definition of the failure criterion for one principle direction.
(σ10 − σ30 ) (σ10 + σ30 )
− sin φ0 − c0 cot φ0 = 0 (2.2)
2 2
where:
φ0 = friction angle
c0 = cohesion
σ10 = major principal effective stress
σ30 = minor principal effective stress
12
The strength parameters c0 and φ0 are not constants, but are dependent on several factors
such as effective stresses, relative density and compaction. For clays the Mohr-Coulomb
criterion is reasonably well applicable, provided that the influence of the pore pressures is
taken into account.
2.2.3 Horizontal soil stresses
The horizontal soil pressure can be expressed as a function of the cohesion and the vertical
soil pressure by means of applying a coefficient for lateral earth pressure K. If the soil is
not subjected to lateral movement and there are no external horizontal forces action on
the soil, the neutral soil pressure is present. The relation for the coefficient of neutral soil
pressure found by Jáky (1948) is often used, which reads:
K0 = (1 − sin φ)OCRsin φ (2.3)
If the soil is subjected to horizontal displacements, the horizontal soil pressures changes. In
case of a retaining wall moving away from the soil, the horizontal pressure reduces. When
the extreme value is reached, this is called the active soil pressure. When the retaining wall
moves towards the soil, the horizontal soil pressure increases until the passive soil pressure
is reached. Several relations have been derived to determine the coefficients for active and
passive soil pressure, for instance by Rankine, Coulomb, Müller-Breslau or Culmann and
Kötter. Primary difference between the methods is the way the slip surface is determined,
and if this sliding plane is either straight or curved.
Stress paths
Stress paths are diagrams that represent the stresses in a specific point by two character-
istic parameters. Important aspect is what parameters to choose for drawing stress paths,
for instance:
• average stress (isotropic stress) 31 (σ1 + σ2 + σ3 ) versus the difference of major and
minor principal stress σ1 − σ3
• mean effective (plane strain) stress 12 (σ1 + σ3 ) versus deviatoric stress 12 (σ1 − σ3 )
If effective stress parameters are used, the stress path is called an effective stress path
(ESP). The effective stress paths for dilatant and contractant material show a different
graph, see Figure 2.1. Here, for both a dilatant and a contractant soil the stress paths are
drawn. The upper blue line represents the shear strength of the soil (Equation 2.1). For
a dilatant soil, the ESP moves to the right which means that the shear strength increases
with increasing normal stresses. For a contractant soil, the ESP bends slightly to the left,
which means that after only a very small increase in normal stress (σ) the shear strength
of the soil is reached.
The distance between effective stress path and total stress path represents the pore pres-
sure. Besides dilatant or contractant behavior, stress paths can be used to visualize many
more features of soil behavior, for example drained soil behavior versus undrained soil
behavior.
13
Figure 2.1: The left figure shows stress paths of a dilatant soil, the right figure shows stress paths of a
contractant soil. (Verruijt, 2012)
2.2.4 Undrained soil behavior

For modelling the behavior of soils and the response of a soil to an external load and/or
displacements or settlements of the soil, a good understanding of the interaction between
soil particles and pore water is essential. When a soil is subjected to compression, the
porosity of the soil decreases and in case of shear the porosity can either decrease (contrac-
tant behavior) or increase (dilatant behavior). In case of a small permeability of the soil,
it will take time for the pore water to adjust to the new situation and reinstate equilibrium
conditions (hydrostatic pore pressures). This process is called consolidation.
Undrained behavior of a soil takes place when there is no sufficient time for the inflow
of water into the pores, or outflow of water from the pores as a reaction to new boundary
conditions. The change in pore pressures will affect the effective soil stresses and therefore
the strength and stiffness of the soil, until the soil is consolidated again.
When only considering the total soil stresses, the strength of the soil can be characterized
by a cohesion only, which is denoted as cu , the undrained shear strength of the soil. As-
suming isotropic effective stress conditions and using the MC failure criterion (Equation
2.2), as a first approximation the undrained shear strength can be estimated by equation
2.4.
σ 0 − σ30 cos φ0 sin φ0
cu = 1 = c0 + σ 0
0 (2.4)
2 1 − 31 sin φ0 1 − 13 sin φ0
where: σ00 = initial mean effective stress
This equation indicates that if φ0 > 0 the undrained shear strength cu increases with
increasing initial mean effective stress σ00 .
However, this formula is derived under the assumption that a volume change can be pro-
duced only by a change of the average effective stress. So anisotropy, dilatancy/contractancy
etc. are not accounted for, and the formula is only a first approximation!
Hydrodynamic period
To quantify if a soil layer responds predominantly in a drained or undrained way, Vermeer
and Meier (1998) have derived the hydrodynamic period T :
kEoed
T = t (2.5)
γ w L2
where:
k = permeability of the soil
14
Eoed = Oedometer stiffness
γw = volumetric weight of pore water
L = drainage length
t = time
Limit values are given for this parameter, indicating drained or undrained behavior con-
sidering one-dimensional consolidation. It is stated that for values of T < 0.01 a loading
situation can be considered undrained. For T > 0.4, sufficient consolidation can take place
and the situation can be modelled as drained. For values of T in between these limits, the
most unfavourable situation should be considered.
In practice this means that the type of soil response is determined by the physical prop-
erties of the soil, but also by the thickness of the soil layer, the duration of the load, and
the speed at which displacements or changes in pore pressures occur.
2.3 Design methods for quay walls

In this section the theoretical background of the design methods is given. First the method
of Blum is explained, which entails a basic, conservative approximation of the minimum
required sheet pile length and the cross-sectional forces in the wall.
Secondly, a more accurate model based on the Euler-Bernoulli beam theory is covered.
Here the retaining wall is modelled as a spring-supported beam where the soil behaves
linear elastic. The software program D-Sheet Piling is used for this method.
Lastly, the approach using the Finite Element Method is described. This method has the
advantage of offering possibilities to describe the behavior of the soil in more detail and
more accurately, using a variety of material models. The software program Plaxis is used
for this method.
2.3.1 Method of Blum

To ensure the horizontal and rotational equilibrium of a sheet pile wall, Blum has devel-
oped an approach in which the required length of the sheet pile can be calculated in an
analytic way. The basic principle of Blums method of analysis is that the sheet pile wall
is considered as fully clamped at its toe, with the additional condition that the bending
moment at the toe is zero. A shear force R can be present at the level of the assumed
clamped edge. It is assumed that during failure the deformations are sufficient to generate
full active or passive soil pressures on each side of the wall.
By analyzing all forces acting on the wall and using the condition that the displacement
of the top of the wall has to be zero (due to an horizontal ground anchor), the required
embedded depth of the wall and thereby the required length of the wall can be calculated.
In Figure 2.2 a schematization of the situation with the acting forces on the wall are
presented.
15
Figure 2.2: Schematization of Blum for a situation with homogeneous soil profile. In the left figure the
internal stresses in the wall are shown. The right figure shows the pressures due to soil and groundwater,
acting on the wall. (Verruijt, 2012)
To enable the generation of the concentrated force R, the wall should be given some ad-
ditional length. This is done by choosing the length of the wall somewhat larger than the
theoretical value computed in the analysis. Schematizing the structure in this way results
in somewhat lower bending moments in the wall, but a larger length of the wall is required.
The Blum calculation is performed using a script which is based on the examples given in
Verruijt (2012). The coefficients of lateral soil pressure are determined using the Müller-
Breslau method, see also 2.2.
Müller-Breslau assumes straight slip surfaces, but also takes into account the wall friction
δ (Deltares, 2017);
cos2 φ0
Ka = r (2.6)
sin φ0 sin(φ0 + δ)

1+
cos δ
cos2 φ0
Kp = r (2.7)
sin φ0 sin(φ0 + δ)

1−
cos δ
with: φ0 = effective friction angle of the soil

δ = friction angle between soil and wall
Modelling undrained soil behavior with the method of Blum

The method of Blum does not take into account undrained soil behavior.
Also, the way Blum’s theory schematizes the soil stresses on the wall neglects certain
features of real soil behavior, i.e. stiffness properties and stress-dependency. This makes
the results only a first estimate of the required length of the wall and forces that are being
generated.
16
2.3.2 Spring-supported elastic beam model
The spring-supported elastic beam model is a design method which enables the implemen-
tation of more complicated boundary conditions. In this model the behavior of the soil
is modelled by uncoupled elasto-plastic springs. The reaction forces of the soil depend on
the properties of the soil, the geometry of the elastic supported beam and the magnitude
of the displacements. The reaction forces of the soil are expressed in terms of force per
meter length, and can be determined using the following relationship:
p=k·w (2.8)
with:
p = soil pressure per unit area [kN/m2 ]
k = modulus of subgrade reaction [kN/m3 ]
w = displacement [m]
For the elasto-plastic analysis of the soil a schematization has to be made of the hori-
zontal soil pressures as function of the displacement of the wall. As already described in
Chapter 2.3.1, the simplistic model of Blum states that the soil immediately behaves as
fully plastic. Based on the direction of the deformation of the wall the soil reaches either
the active or passive soil pressure. The spring-supported elastic beam model however,
states that the soil reaches the active or passive soil pressure only after a certain displace-
ment takes place. In between these two extreme values, the soil behaves linear elastic and
the soil pressure depends on the magnitude of the displacement of the wall. When no
displacement has occurred yet, the soil is in the neutral state and the horizontal stresses
in the soil equals the neutral soil pressure.
The software D-Sheet Piling calculates the soil pressures for many points along the wall.
Because the soil pressures, displacements, water pressures and internal forces in the wall
are all correlated, the calculations are performed in an iterative manner.
Also, multiple construction stages can be defined, taking into account the stress and strain
history from previous steps. The effects of creep and arching of the soil are neglected
(CUR166, 2008).
The coefficients of active and passive soil pressure can be determined with several meth-
ods. In D-Sheet, three approaches are available: the method of Culmann and method of
Müller-Breslau which both use straight sliding planes, and the method of Kötter which
uses a curved sliding plane. It is also possible to enter the coefficients, and other parame-
ters, manually. An example of the interface of D-Sheet is given in Figure 2.3.
17
Figure 2.3: Example of a retaining wall in D-Sheet Piling
Undrained soil behavior in D-Sheet Piling

D-Sheet Piling does not take into account undrained conditions and undrained soil be-
havior. The undrained shear strength could be determined by stating that in undrained
conditions φ0 = 0 and the undrained cohesion cu represents the undrained shear strength.
The analysis, in which the friction of the material and the pore pressures are neglected, is
called an undrained analysis. In general the following relation holds:
1
su = (σ1 − σ3 ) (2.9)
2
When assuming that in a saturated soil there can be practically no volume change in
undrained conditions, the isotropic effective stress remains constant. In case of failure,
the major and minor principle stresses σ10 and σ30 must satisfy the Mohr-Coulomb failure
criterion. Combining this information with Equation 2.9 results in the following relation
for the undrained shear strength (Verruijt, 2012):
σ10 − σ30 cos φ sin φ
su = =c 1 + σ00 (2.10)
2 1 − 3 sin φ 1 − 31 sin φ
This formula does not take into account the effects of anisotrophy and dilatancy, and is
therefore only a first approximation. Also, when the situation occurs that the excess pore
pressures in the soil can dissipate without a decrease in loading, this relation is not valid
anymore.
2.3.3 FEM - Plaxis

Plaxis is a software package that uses a finite element method that enables the analysis
of deformation, stability and groundwater flow in geotechnical engineering. It is widely
18
used in practice for a variety of design and engineering problems. The software offers a
variety of models to describe the mechanical behavior of soils and rocks, both very basic
models which serve as crude approximations, but also more advanced models which take
into account essential properties of soil behavior such as stress-dependency of stiffness,
anisotropy and time dependencies.
The Plaxis software will divide the model space into a number of elements (mesh). For each
element, both stresses and strains are calculated using predefined constitutive relations.
These relations differ for each material model that can be assigned to a certain soil layer
in the program. The material models which are used in this research, along with their
essential properties, are briefly covered in section 2.5.
19
2.4 Developments in dike safety assessment
In this section the main properties and characteristics of a dike are presented, and how
this relates to the properties of a quay wall. Also, the new design approach for dikes is
explained.
2.4.1 Design process for a dike vs quay wall

The primary function of a dike is to retain water. A dike should therefore be sufficiently
high, and also be able to withstand the horizontal water pressures and possible wave
attacks. Figure 2.4 shows an example of a typical dike with its most important elements.
Figure 2.4: General overview of dike showing the most important elements. Source: Jonkman et al.
(2018)
Due to the difference in water levels between the inner and outer side of the dike, seepage
through (and underneath) the dike will occur. To prevent this, the dike body should
be constructed to be sufficiently impermeable for the water to flow through. Often the
dike core or dike cover is constructed of clayey soil with a low permeability. This will
prevent seepage and possible outflow of soil from the dike. However, due to the increased
hydrostatic pressure acting on the dike at times of high water, the pore water pressures
inside the dike core and dike cover will increase as well, which results in a decrease of the
effective soil pressure. Most importantly, a lower effective soil pressure results in a lower
shear strength of the soil (see equation 2.1) and hence can result in slope instability.
This effect can be countered in several ways, the most commonly used is increasing the
dike width, i.e. reducing the slope steepness of the dike. At places with limited available
space other strength increasing measures are applied, such as inserting stability screens
inside the dike body. Figure 2.5 gives an example of a sheet pile wall constructed inside
the dike body, connected with grout anchors.
20
Figure 2.5: Example of a dike with a stability screen. In this case there are grout anchors applied along
the screen, to provide more horizontal support if the soil body starts shearing. Source: POVM
The assessment regarding strength and stability of a dike with stability screen shows some
similarities with that of a quay wall. Most important, in both cases there are one or more
structural elements situated inside the geotechnical body to prevent large deformations
and sliding of the soil, in the case the shear resistance of the soil is exceeded. However,
the new design approach is developed for the specific situation where there is a high water
level on one side of the dike. So the external load is caused by a horizontal (hydrostatic)
water pressure. In case of a quay wall, the external load working on the wall is caused by a
difference in retaining height in combination with surcharge loads. This has consequences
for the development of soil stresses when the structure is experiencing the expected loads.
During a high water wave the dike is loaded by horizontal pressures in the form of a hy-
drostatic water table. The dominant phenomenon here is thus that the total soil pressures
are increased due to an increase in water pressures. Because of the low permeability of
the dike body this will result in excess pore pressures.
For the quay wall holds that the total soil pressure increases due to additional surcharge at
the surface level. Difference here is that the surcharge acts as effective vertical soil pressure
at the surface level of the retaining side of the wall. However, if the permeability of the
soil is relatively small compared to the load duration or speed of loading, the pore water
pressures cannot follow the new stress situation and excess / deficit pore water pressures
will occur.
2.4.2 New design approach for dikes based on critical state soil mechan-
ics
With respect to the previous assessment round of primary flood defences several changes
have been implemented regarding the modelling of soil behavior. The most important
changes are (Rijkswaterstaat, 2017):
1. the switch of the Mohr-Coulomb model to Critical State Soil Mechanics theory for
describing the soil behavior.
2. differentiate between drained and undrained behavior of the soil during failure and
using the SHANSEP method to determine the undrained shear strength during the
critical loading situation. Section 2.5 will give more information about how the
undrained behavior of the soil is modelled with this new approach.
21
Introduction of cssm framework
The critical state is defined as the density state where the granular materials shear at
a constant volume and stress state. At this state, the dilatancy rate diminishes and
approaches zero. The critical state framework determines the shear strength of the soil
based on the ultimate strength instead of the peak strength of the soil, distinguishing
between normally consolidated and overconsolidated soils (Rijkswaterstaat, 2017).
In the normally consolidated phase of the soil, the failure envelope is described using the
following formula:
tmax = s0 sin φ0cv (2.11)
where:
tmax the maximum mobilized shear strength,
1
s0 the mean effective stress (σv + σh ),
2
φcv the critical state friction angle
The shear strength of the soil is based on the mean effective stress in the soil (s0 ) and the
critical state friction angle φcv . Connecting the values of tmax for different stress states
will give a line with slope sin φ0cv . This line represents the critical state line and con-
nects the peaks of the Mohr’s circles at failure of the soil. See Figure 2.6 for a graphical
representation. In this approach the cohesion is not explicitly taken into account, but is
incorporated in the over consolidation ratio (OCR).
Figure 2.6: Shear strength according to CSSM. (Rijkswaterstaat, 2017)
In the normally consolidated state of the soil, there is no cohesion. In the overconsolidated
state, the cohesion provides the additional strength resulting in the peak strength. The
behavior of soil during undrained loading situations is determined by the local conditions
22
of the soil. These are expressed in terms of the yield stress σvy and the OCR.
Next to the critical state framework, also the soil behavior based on an undrained response
is introduced for the assessment of slope stability. The primary and most essential change is
the use of the NGI-ADP model Grimstad (2011). This model aims to predict and describe
the undrained soil behavior based on undrained strength and stiffness parameters, and
taking into account anisotropy of the soil. For the detailed calculations at the stage near
failure, the model is adjusted to determine a new stress state based on the SHANSEP
theory. More information about these models can be found in section 2.5.
2.5 Material models for describing soil behavior

In this section a number of models are described which are used in the FEM software
PLAXIS to model essential features of soil behavior. Each model has different character-
istics and has its own advantages and limitations for various applications.
2.5.1 Mohr-Coulomb
The Mohr-Coulomb soil model describes the soil behavior as linear elastic perfectly plas-
tic and can be used to obtain a first estimate of the deformations. Because it is perfectly
plastic, there is no hardening or softening of the soil taken into account in this model.
Mohr-Coulomb can not model the critical state of the soil, because the dilatancy continues
forever during plastic shearing.
In the Mohr-Coulomb model, undrained behavior could be modelled by setting the friction
angle φ equal to zero and the cohesion c to cu , the undrained shear strength. However,
when using this approach only the undrained shear strength is taken into account. Re-
garding the stiffness of the soil, still the drained situation is assumed. Also, this model
does not take into account the process of consolidation, and therefore with this approach
caution is required for loads of longer duration (Plaxis, 2018).
2.5.2 Hardening Soil

The Hardening Soil (HS) material model takes into account the stress dependent stiffness
behavior of soils, by using a power-law formulation for the stiffness parameters. It also
has multiple stiffness parameters: for primary loading the secant stiffness E50 and the
unloading/reloading stiffness Eur . These properties are visualized in Figure 2.7.
HS takes into account the compression hardening and the shear hardening of the soil.
Compaction hardening is used to model plastic strains that are generated during primary
compression. The shear hardening includes the generation of plastic deviatoric strains
by mobilizing the materials internal friction. During this process the stiffness of the soil
slowly decreases.
23
Figure 2.7: Hyperbolic stress-strain relation of sand in primary loading for a standard drained triaxial
test. (Plaxis, 2018)
In 2.7 it is shown that the Hardening Soil model takes into account the stress- and strain
dependency of the stiffness. The initial stiffness Ei for instance has a larger value than
the secant stiffness E50 at 50% of the maximum strength of the soil. The Hardening
Soil small strains (HSss) material model is particularly suited for situations where small
strains occur. The application of cssm however, considers the critical state where very
large strains are present. Therefore the HSss model is not taken into account in this
study.
2.5.3 Soft Soil

The Soft Soil (SS) material model may be used for near normally consolidated clay-type
soils (Plaxis, 2018). This class of materials is best characterized by their high degree
of compressibility. Compared to normally consolidated sands, the normally consolidated
clays behave up to ten times softer. For these extreme compressibility the Soft Soil model
is suitable.
This model takes into account the stress-dependency of soil stiffness, distinction between
primary loading and unloading-reloading and uses the Mohr-Coulomb failure criterion,
similar to the Hardening Soil model. Important difference is the logarithmic dependency
between stresses and strains, compared to the hyperbolic relation in the HS model.
2.5.4 NGI-ADP
The NGI-APD model may be used for bearing capacity, deformation and soil-structure in-
teraction analyses involving undrained loading of clay (Plaxis, 2018). Distinction is made
between three different stress states; compression (Active mode), neutral (Direct Simple
Shear mode) and extension (Passive mode). For each mode, both the undrained shear
24
strength and the shear strain at failure are defined. This way the model takes into ac-
count the anisotrophy of undrained shear strength and stiffness of the soil.
Figure 2.8: Stress-strain relation of clay and peat. The three curves represent the results of an undrained
triaxial compression test (sA P
u ), an undrained triaxial extension test (su ) and a Direct Simple Shear test
(sDSS
u ). Gur defines the isotropic elasticity of the soil. (Naves and Lengkeek, 2017)
A list of the model variables of the NGI-ADP model is given in Table 2.1.
Symbol Description
G/su A Ratio unloading/reloading shear modulus over (plane strain) active shear strength
γf C Shear strain at failure in triaxial compression
γf E Shear strain at failure in triaxial extension
γf DSS Shear strain at failure in direct simple shear
su A Reference (plane strain) active shear strength
ref
vertref Reference depth
su A Increase of shear strength with depth
inc
su /su A
P Ratio of (plane strain) passive shear strength over (plane strain) active shear strength
τ0 /su A Initial mobilized shear resistance
su DSS /su A Ratio of direct simple shear strength over (plane strain) active shear strength
ν Poisson’s ratio
νu Undrained Poisson’s ratio
Table 2.1: Shansep NGI-ADP model variables. (Brinkgreve and Panagoulias, 2017)
25
2.5.5 Shansep NGI-ADP
The Shansep NGI-ADP material model is a user defined material model which combines
the NGI-ADP material model with the SHANSEP (Stress History And Normalized Soil
Engineering Properties) concept. The advantage of this model is that it is able to simulate
potential changes of the undrained shear strength based on the effective stress state of the
soil (Brinkgreve and Panagoulias, 2017).
The undrained shear strength of the soil is determined using the SHANSEP method (Ladd,
1974):
0
su = σvi · S · OCRm (2.12)
S and m are normalized soil parameters. The variable S determines how fast and to which
extend the stress path bends away from the vertical to the left, and reaches the CSL.
The power m determines to what extend the effect of the load history (σvy 0 and OCR)
influences the undrained shear strength. The value of m can be between 0.5 and 1.0 where
m = 1.0 represents linear elastic soil behavior.
The over consolidation ratio OCR depends on the site characteristics, actual stresses and
loading history.
0
σvy
OCR = 0 (2.13)
σvi
where:
0
σvy yield stress,
0
σvi in-situ effective stress.
This material model is developed such that at first it models the soil using the properties
of the NGI-ADP model. Only after the user activates a switch, the Shansep - MC material
model is being activated and the new soil stresses are calculated according the relation
given in equation 2.12. The value of σvi 0 in equation 2.12 is being extracted from the
output of the last calculation step, and the yield stress (or pre-consolidation stress) is the
highest stress level the soil has experienced until that moment.
26
CHAPTER 3
Method
This chapter describes the procedure that has been used to elaborate the case studies.
This includes the specific applications of the different design methods for the case studies,
important modelling choices and assumptions that have been made.
3.1 Framework case studies

To help answer the main research question ”Can critical state soil mechanics in combina-
tion with undrained analysis of the soil be applied for the design of quay walls, and what
are the effects for the prediction of strength and stability? ”, three different case studies of
a quay wall are elaborated. The primary difference between the case studies is the soil
profile surrounding the quay wall. All other variables such as different loading types, the
retaining height and the geometry of the quay wall, the strength and stiffness of the struc-
tural elements, and the construction phases are all kept constant across the case studies.
Each case study covers a different type of soil profile for the Holocene part of the soil,
namely:
• Case study 1: soil profile consisting of mainly sandy soils.
• Case study 2: soil profile where the Holocene part of the subsoil contains a nor-
mal consolidated clay layer. In this case study the effects of the alternative design
method on soft soils are investigated. The anisotrophy of the soil along the sliding
plane will be taken into account in this case study.
• Case study 3: soil profile where the Holocene part of the subsoil contains an over-
consolidated clay layer. In this case study, both the anisotrophy of the soil along the
sliding plane and the stress history of the clay layer are taken into account.
An overview of the three case studies is presented in Figures 3.1 up to Figure 3.3. The
figures are adapted from the Plaxis user interface.
27
Figure 3.1: Overview of soil layers and soil types for case study 1.
28
For each case study, two different design calculations will be performed.
1. Design calculations based on the conventional design approach. For the detailed
calculations D-Sheet or Plaxis are used, in accordance with the guidelines presented
by CUR166 (2008) and CUR211 (2014).
2. Design calculations based on the alternative design approach. Here the Plaxis
material model NGI-ADP Shansep model is recommended by POVM when ana-
lyzing a stability screen inside a dike body. Therefore this material model will be
applied in the plaxis calculations, when performing an undrained analysis of a quay
wall using the alternative design approach.
3.1.1 General modelling choices

Characteristic values versus mean values
For the alternative design approach, no partial safety factors are deduced for quay walls
so far. Therefore, in this research the mean value for each of the design parameters will
be used for both the conventional and alternative design approach. This approach ensures
that the input used for the calculations is identical.
The results of the design calculations are quantified using the maximum displacements of
the quay wall and the sectional forces of the quay wall. A comparison is made between
the values of these quantities found using both design approaches. The relative difference
between the results will give an indication whether the alternative design approach shows
any promising results.
29
Software used for the conventional design approach
The results of the conventional design approach are generated using the software programs
D-Sheet Piling and Plaxis. For D-Sheet Piling, the strength parameters are obtained
from triaxial compression tests, and the stiffness parameters are obtained from design
calculations of the reference project.
For the plaxis calculations, the material model Hardening Soil is used for both sand and
clay layers. Specific correlations that are used to determine the model parameters are
given in appendix B.
Also, a preliminary calculation is performed using the method of Blum. The result of this
calculation is used as starting point for the detailed design.
Software used for the alternative design approach

For the alternative design approach only calculations in Plaxis are performed. Essential
for this approach is the modelling of undrained soil behavior. Both D-Sheet Piling and
the method of Blum do not have the ability to perform this.
For every soil layer, an analysis is made whether an undrained response of the soil is
expected. In that case, soil layers below the phreatic surface consisting of clay are modelled
using the NGI-ADP-Shansep material model. Sand layers as well as clay layers above
the phreatic surface have to be modelled using a drained analysis where the strength
parameters at the critical state are used. (POVM BEEM, 2018). For the calculations,
different construction phases have been defined.
An overview of these phases is presented in Figure 3.4
Figure 3.4: Overview of construction phases used in the calculations
30
3.1.2 Specific modelling choices per case
Case study 1 - quay wall Amazonehaven
For the first case study, the Amazonehaven quay wall is used as reference project. This
project provides a large set of soil investigation data, technical drawings and structural
properties of the quay wall.
The results of the calculations performed by Blum & D-Sheet Piling are used as benchmark
values and will indicate whether the Plaxis calculations are reasonably accurate.
Since this case study does not include undrained analysis of the soil, both conventional
and alternative design methods use effective strength and stiffness parameters. For the
alternative design approach, new strength parameters are derived to represent the critical
state of the soil. More detailed information about parameter derivation for the critical
state is given in Appendix A.1.
Case study 2 - quay wall Alblasserdam Container Terminal

In this case study the development of the Alblasserdam Container Terminal is used as
reference project. This project provides a set of soil investigation data that consists among
others information about soft soils, in particular normal consolidated clay. The available
information of these soils is used to construct an alternative soil profile for the quay wall
structure. In this profile a large clay layer with the aforementioned properties is present.
The structural properties of the quay wall, geometry and loading types are identical to
those of case study 1.
Case study 3 - quay wall Groningen

In this case study the development of a deep foundation pit in the province of Groningen
is used as reference project. Characteristic for this location is the presence of boulder clay,
large layers of fine clay which are overconsolidated as a result of large ice loads in the past.
Again, the available soil investigation data is used to create a representative soil profile.
The structural properties of the quay wall, geometry and loading types are identical to
those of case study 1.
The similarities and differences of modelling choices and boundary conditions between the
case studies are summarized in Table 7.1.
Soil types Retaining height [m] User load [kN/m2 ] Wall type
Case 1 Sand 17.55 20-70 1524*19 + 3PU28
Case 2 NC clay & sand 17.55 20-70 1524*19 + 3PU28
Case 3 OC clay & sand 17.55 20-70 1524*19 + 3PU28
Table 3.1: Similarities and differences of modelling choices and boundary conditions for the case studies
31
3.2 Framework for determining soil parameters
In this section background information is provided about the various soil parameters used
in the analysis. The most important parameters for each design method are briefly ex-
emplified, and it is explained how these parameters can be derived from soil investigation
data.
3.2.1 General soil parameters

Friction angle
The friction angle φ represents the amount of friction between the soil particles. Together
with the cohesion it determined the shear strength of the soil, according to Equation 2.1.
To determine its value, triaxial tests (drained for sand, undrained for clay) or direct simple
shear tests (for peat) are used. In general, φ is related to the peak strength of the soil, at
2% axial strain.
For the application of the new design approach, the critical state friction angle φcv is used.
The standards for the new design approach for dikes prescribe the following conditions
and set up for triaxial tests (Rijkswaterstaat, 2017):
• Anisotropic consolidation on either undisturbed or prepared soil specimen.
• The prepared soil specimen should have a void ratio close to the in-situ value of the
soil.
• The consolidation stress is chosen to be higher than the preconsolidation pressure
or equal to the in situ vertical effective stress, so that yielding of the soil is assured.
• The specimen has to be sheared up to at least 25% axial strain in case of sand or
clay, and to at least 40% axial strain in case of peat. These strain levels ensure that
it is safe to assume that the critical state of the soil is reached.
Cohesion
Cohesion c encloses the Van der Waals-forces, the effect of cementation and the capillary
forces in (partly) saturated soils. The value for the cohesion is related to the peak strength
of the soil, at 2% axial strain. In the critical state at normally consolidated state, the
cohesion of the soil is assumed to be zero.
Dilatancy angle
The dilatancy angle ψ quantifies the amount of volume expansion during shearing of the
soil. For normally consolidated clay and for peat, the dilatancy angle is equal to zero. For
quartz sand it can be specified using the following relation (Plaxis, 2018):
ψ = φ0 − 30° (3.1)
Undrained shear strength

The undrained shear strength quantifies the resistance of the soil during shearing at
undrained conditions. Values can be obtained directly from undrained triaxial tests, or
using correlations. Relation between undrained shear strength and cone resistance of CPT:
su = qnet /Nkt (3.2)
32
Here qnet represents the cone resistance, corrected for pore water pressure effects and total
stresses. Nkt is an empirical factor, generally around 20.
Yield stress
The last parameter that is important for cssm is the yield stress σvy0 , a measure for the
loading history of the soil. The relation between the yield stress and the in-situ effective
vertical stress can be presented in two ways:
• σvy
0 = σ 0 + P OP
vi
• σvy
0 = σ 0 · OCR
vi
where POP is the pre-overburden pressure.
3.2.2 Shansep model parameters

Undrained shear strength ratio S
The undrained shear strength ratio S gives an indication of the magnitude of the undrained
shear strength in relation to the stress state of the soil. The parameter S has to be
determined using the relation:
qcv
S= (3.3)
σvc
The soil specimen has to be consolidated undrained triaxial test, using a consolidation
stress σvc . This value corresponds with the stress at the point in a sigma-epsilon diagram
of a CRS test, where the tangent of the primary loading branch first touches the settlement
curve. This point is represented by arrow B in Figure 3.5.
33
Figure 3.5: Stress-strain diagram of an CRS test. The blue line represents the test results. The red line
through points A and B represents the tangent line of the primary loading branch of the diagram. Source:
Rijkswaterstaat (2017)
Here qcv is the deviator stress at constant volume of the soil, and follows from the triaxial
test mentioned above. In case of overconsolidated soil, S can be determined using the same
approach where the soil specimen is now consolidated at the estimated in-situ vertical
effective stress σvi . Overall, the value of S increases with decreasing volumetric weight of
the soil.
Power coefficient
The power m determines to what extend the stress history of the soil (σvy 0 ) influences the
undrained shear strength. The value of m can be between 0.5 and 1.0 where m = 1.0
represents linear elastic soil behavior. Typical values are between 0.7 and 0.8.
To determine this parameter, the following relationship can be used;
b−a
m= (3.4)
b
where a and b are the isotachen parameters from an oedometer test or CRS test.
3.2.3 NGI-ADP model parameters

Parameter determination according to Post and Luijendijk (2018) and (POVM BEEM,
2018). The following correlations and default values have been used for the calculations
34
using the Shansep NGI-ADP material model;
sin φcv
S≈ (3.5)
2
Cs
m≈1− (3.6)
Cc
Cc and Cs are the primary and secondary compression indices. They take into account
the effect of virgin loading and creep on the one-dimensional compression of a soil.
su P /su A = e−2β (3.7)
√
3η0 Λ
β= (3.8)
2M
3(1 − K0 )
η0 = (3.9)
1 + 2K0
Λ=m (3.10)
6 sin φ0
M= (3.11)
3 − sin φ0
1
su DSS /su A ≈ · (1 + su P /su A ) (3.12)
2
u
E50
G≈ (3.13)
3
3
γfC = εC (3.14)
2 1
Parameter Symbol Value Unit

Reference interface stiffness Eoed ref 15.0 E3 kN/m2
Reference interface cohesion c0ref 1.0 kN/m2
Friction angle φ 8.0 °
Dilatancy angle ψ 0 °
Power coefficient determining stress-dependency of stiffness UD-Power 1.0 -
Reference stress level UD-P ref 100 kN/m2
Table 3.2: Strength and stiffness parameters of the interface of boulder clay for the Shansep NGI-ADP
model
35
CHAPTER 4
Case study 1 - Quay wall Amazonehaven
In this chapter the first case study is elaborated. In section 4.1 a brief, general introduction
of the case study is given. Specific characteristics for this case study and several boundary
conditions and technical specifications are given in section 4.2. The designs based on the
conventional design approach and the new design approach are presented in sections 4.4
and 4.5. Finally, section 4.6 gives an overview of the results of both designs and analysis
of the outcome.
4.1 Introduction
The first case study includes the quay wall of the Amazonehaven in the Port of Rotterdam.
This quay wall structure is constructed in 2011-2012, and has replaced the existing quay
wall to create a larger entrance of the Amazonehaven harbour branch. In Figure 4.1 the
location of the Amazonehaven in the Port of Rotterdam is given.
36
Figure 4.1: Location of the Amazonehaven in the Port of Rotterdam. Source: Google Inc.
The quay wall is designed using several sections, each having different dimensions based
on their specific requirements and boundary conditions. Taking into account the available
data and information from CPT’s and lab tests of the different sections, the first part of
the quay wall (Kade A1 in Figure 4.2) is used for this research. This quay wall and area
behind it serve for the transshipment and storage of coal.
Figure 4.2: Overview Amazonehaven with different quay walls. ’Kade A1’ and ’Kade A2’ are indicated
with the solid green respectively purple line. The dotted yellow line indicates the old quay wall, which has
been demolished afterwards. Source: Gemeentewerken Rotterdam (2011a)
37
The soil profile consists of mainly sandy soils with good permeability. Therefore for this
case study the new design approach is only partially applied. The critical state approach
is used for the design of this quay wall, but the undrained analysis is not. This new design
is further elaborated in the following sections.
Primary sources of data used in the design calculations are the (geo-)technical reports
of BAM Infra (2009), Fugro GeoServices B.V. (2011) and Gemeentewerken Rotterdam
(2011a).
38
4.2 Parameters / Boundary conditions
Quay wall structure
The quay structure consists of a combined wall of tubular piles with interconnecting sheet
piles. On top of the combined wall a large capping beam of reinforced concrete is con-
structed. This beam will distribute the forces over the wall, and also provides support for
the mooring bollards. The beam has a width of 2.5m and a height of 5.0m.
The wall is anchored by two rows of grout anchors connected to each tubular pile. The
tubular piles are the main structural elements and transfer the loads induced by the ships,
the soil and additional surcharge loads to the subsoil. The piles are placed every 3.38m
in longitudinal direction of the quay wall. A cross-section of the quay wall including some
dimensions is given in Figure 4.3.
Figure 4.3: Cross section of quay wall A1. Source: Gemeentewerken Rotterdam
Structural and geometric properties

The following list shows the characteristic values used for modelling the quay wall. All
structural parameters are averaged per meter length of the wall.
• Final surface level behind the quay wall: NAP + 5.0m
• Embedded depth of tubular piles: NAP - 31.0m
• Embedded depth of sheet pile elements between the tubular piles: NAP - 16.55 m
• Normative depth of the bottom in front of the wall: NAP - 12.55m
39
• Normative depth at center of harbour branch Amazonehaven: NAP - 18.15m
• Normative low water level: NAP - 0.68m
• Normative hydraulic head behind quay wall, in case of poor drainage: NAP + 0.59m
• Stiffness of combined wall: EI = 1.649 · 106 [kNm2 /m]
• Stiffness of tubular piles: EI = 1.579 · 106 [kNm2 /m]
• Young’s modulus steel: E = 2.1 · 105 [=N/mm2 ]
• Anchor 1: angle 40°, length 47m of which the last 12m consists of the grout body.
• Anchor 2: angle 45°, length 44m of which the last 12m consists of the grout body.
• Anchor bar cross section 1.59 · 10−3 [m2 /m], design yield force 896 [kN/m]
• Grout body diameter 0.38m.
• Axial skin resistance of the anchor: 400kN
• Both wall and anchor have elastoplastic material behavior
Loads
During the lifetime of the structure, it is subjected to several external loads. The loads
that result in the normative loading situation are listed below:
• Vertical point load representing the weight of the capping beam: 270 [kN/m]
• Distributed surface load: 20 - 70 [kN/m2 ], see Figure 4.4
• Pre-tension in anchors: 350 [kN/m]
Figure 4.4: Schematization of the variable load behind the quay wall
4.2.1 Soil profile

Along the existing quay wall and around the center line of the newly build quay wall,
multiple CPT’s and boreholes have been executed. Based on the data from the soil in-
vestigations, a soil profile has been constructed. This geotechnical length profile varies
along the quay wall. For each section of the quay wall a normative soil profile has been
determined which is representative for the entire section.
For this report, part 1 of section A1 of the quay wall is used. The geotechnical length
profile of this section, based on both recent and historical soil investigation data plus en-
40
gineering experience in Rotterdam, is presented in Figure 4.5.
Figure 4.5: Geotechnical length profile of the Amazonehaven quay wall part A1. Source: Gemeentewerken
Rotterdam
The normative soil profile that is used for the design calculations is presented in Table 4.1.
The layer numbers correspond to those of Figure 4.5. The values for the soil parameters
γ, φ and c are used as input parameters for the various design calculations.
41
Top of layer γdry /γsat C
Layer Soil type φ [°]
[m NAP] [kN/m3 ] [kPa]
1 +5.0 Sand, anthropogenic 17.2 / 19.2 30 0
Sand, beach sand
2 -1.5 17.4 / 19.4 32.5 0
formation
3 -2.7 Sand, beachsand, low qc 17.2 / 19.2 30.0 0
Sand, beach sand
2 -4.7 17.4 / 19.4 32.5 0
formation
Sand, ’wadzand’ formation,
5 -5.5 17.2 / 19.2 30.0 0
with few thin clay layers
Sand, ’wadzand’ formation,
7 -11.8 17.2 / 19.2 32.5 0
high qc
Sand, deep ’wadzand’
8.1 -13.3 formation, with many thin 17.0 / 19.0 27.5 0
clay layers
8.2 -16.3 17.2 / 19.2 30.0 0
formation
8.1 -17.5 formation, with many thin 17.0 / 19.0 27.5 0
clay layers
Clay, occasionally thin
9 -19.6 15.0 / 15.0 17.5 10
peat layers
10 -21.2 Sand, Pleistocene 18.0 / 20.0 32.5 0
11 -29.5 Silty clay / clay 17.0 / 17.0 22.5 5
10 -30.0 Sand, Pleistocene 18.0 / 20.0 32.5 0
Table 4.1: Normative soil profile and parameters of quay wall part A1, profile1. In the 3rd column the
volumetric weights of the soil in both dry and saturated state are given. The 4th and 5th column give the
friction angle and cohesion of the soil for an axial strain level of 2%, based on historical data & triaxial
tests.
Model parameters Symbol Value Unit

Axial stiffness of combined wall EA1 4.05E+06 kN/m
Bending stiffness of combined wall EI1 1.65E+06 kN/m2 /m
Moment capacity combined wall Mp,wall 5076 kNm/m
Bending stiffness of tubular piles EI2 1.579E+06 kN/m2 /m
Moment capacity of tubular piles Mp,tube 4031 kNm/m
Axial capacity of tubular piles Np,tube 1638 kN/m
Table 4.2: Structural parameters of quay wall used in Plaxis models
42
center to center distance Lspacing 3.38 m
Axial stiffness bar EAbar 1.129E+06 kN
Axial capacity in tension Fmax,tens 3028 kN
Young’s modulus of grout body Egrout 7.07E+06 kN/m2
Diameter grout body Dgrout 0.38 m
Plastic moment capacity anchor Mp 1000 kNm/m
Plastic axial capacity anchor Np 4000 kN/m
Axial skin resistance - Linear -
Skin resistance start Tskin,start,max 286.5 kN/m
Skin resistance end Tskin,end,max 286.5 kN/m
Table 4.3: Structural parameters of grout anchor used in Plaxis models
43
4.3 Design verification calculation based on theory of Blum
In this section the calculation based on the theory of Blum is presented. Blum is used
to check the required embedding depth of the wall to ensure stability of the wall. It is
assumed that all soil layers show a drained response. Therefore the effective strength
parameters φ0 and c0 are used. As already discussed in chapter 2.3.1, the coefficients of
lateral earth pressure are determined using the theory of Müller-Breslau.
Because the first 5 soil layers (from NAP + 5.0m till NAP - 13.3m) have almost the same
volumetric weights and strength parameters, they are combined to one representative layer
to simplify the calculation.
Parameters
The input parameters used for the calculation are given in Table 4.4.
Top of layer Bottom of γdry /γsat

φ0 [°] c0 [kPa] δ [°]
[m NAP] layer [m NAP] [kN/m3 ]
+5.0 -13.3 17.2 / 19.2 30 0 20
-13.3 -16.3 17 / 19 27.5 0 18.3
-16.3 -17.5 17.2 / 19.2 30 0 20
-17.5 -19.6 17 / 19 27.5 0 18.3
-19.6 -21.2 15 / 15 17.5 10 5.83
-21.2 -29.5 18 / 20 30.0 0 16.6
-29.5 -30.0 17 / 17 22.5 5 7.5
-30.0 -35.0 18 / 20 30.0 0 16.6
Table 4.4: Input parameters used to perform Blum calculation
Some additional factors to be taken into account for the calculation:

• the anchor is modelled as if it makes an angle of 90°with the wall. Therefore the
resulting anchor force has been increased to correct for the actual angle of 45°.
• the coefficients of lateral soil pressure are determined using the theory of Müller-
Breslau, which is valid for friction angles φ0 ≤ 30° (see Ch28.2 of D-Sheet manual).
This results in a small underestimation of the passive soil strength for the Pleistocene
sand layer.
• The top load as presented in Figure 4.4 is for the calculation represented as a uniform
load of 20 kN/m2 , because Blum’s theory looks only at the soil stresses right next
to the wall.
• No leveling of pore pressures due to different water table at left/right side of wall
Results
The results of the calculations are presented in Table 4.5. The calculated anchor force and
bending moments are later compared with the results found by D-Sheet and Plaxis.
44
Variable Value
Anchor Force 2238 kN
Mmax 1583 kNm/m1
Required embedding depth wall 24.8 m - NAP
Table 4.5: Results of Blum calculation
4.4 Designs based on conventional design approach

Design using D-Sheet Piling
In this section the design of the Amazonehaven quay wall, based on the current design
practices, is presented. For this design, a model of the quay wall and soil profile is build
in D-Sheet Piling. The structural properties presented in Chapter 4.2 and the soil profile
given in Table 4.1 have been used as input for the calculations. This way the calculations
coincide with those performed for the actual design of the Amazonehaven. For the existing
design calculations, the tangent option (1 branch) in D-Sheet has been used to describe
the stiffness of the soil. Since only 1 curve for the spring characteristic is used, the soil
behavior is assumed to have a bi-linear stress-strain relationship.
A more accurate result can be obtained using the secant option, which takes into account
the stress-dependency of the soil stiffness. Assuming that the soil properties of the differ-
ent soil layers are constant along the quay wall, the values for the modulus of subgrade
reaction have been derived using the report of the design for cross section A3 (see Figure
4.2). The values found are presented in Table 4.6.
k1 k2 k3
Layer nr.
[kN/m2 /m1 ] [kN/m2 /m1 ] [kN/m2 /m1 ]
1 16000 8000 4000
2 30000 15000 7500
3 16000 8000 4000
2 30000 15000 7500
5 16000 8000 4000
7 20000 10000 5000
8.1 12000 6000 4000
8.2 16000 8000 4000
8.1 12000 6000 3000
9 2000 800 500
10 40000 20000 10000
11 6000 4000 2000
10 40000 20000 10000
Table 4.6: Overview of the values for modulus of subgrade reaction using the secant method. Based on
geotechnical report of Amazonehaven quay wall section A3
The results of the calculations are presented in Table 4.9, at the end of this section. More
detailed information about the design calculations is presented in Appendix C.
45
Design using Plaxis
In this section the design of the Amazonehaven quay wall is presented using a Finite
Element Method. A model of the quay wall and soil profile has been made in Plaxis.
Again, the structural properties presented in Chapter 4.2 and the soil profile given in
Table 4.1 are used as input for the model. This has resulted in the following list of input
parameters:
Top of layer Material γdry /γsat C

Layer φ0 [°] ψ [°] Rinter
[m NAP] model [kN/m3 ] [kPa]
1 +5.0 Sand 17.2 / 19.2 30 0 0 0.66
2 -1.5 Sand 17.4 / 19.4 32.5 0 2.5 0.66
3 -2.7 Sand 17.2 / 19.2 30.0 0 0 0.66
2 -4.7 Sand 17.4 / 19.4 32.5 0 2.5 0.66
5 -5.5 Sand 17.2 / 19.2 30.0 0 0 0.66
7 -11.8 Sand 17.2 / 19.2 32.5 0 2.5 0.66
8.1 -13.3 Sand 17.0 / 19.0 27.5 0 0 0.66
8.2 -16.3 Sand 17.2 / 19.2 30.0 0 0 0.66
8.1 -17.5 Sand 17.0 / 19.0 27.5 0 0 0.66
9 -19.6 Clay 15.0 / 15.0 17.5 10 0 0.5
10 -21.2 Sand 18.0 / 20.0 32.5 0 2.5 0.66
11 -29.5 Clay 17.0 / 17.0 22.5 5 0 0.5
10 -30.0 Sand 18.0 / 20.0 32.5 0 2.5 0.66
Table 4.7: Normative soil profile and strength parameters inserted in Plaxis.
Top of layer ref ref

Layer E50 ref Eoed Eur m
[m NAP]
1 +5.0 25 25 100 0.5
2 -1.5 30 30 120 0.5
3 -2.7 25 25 100 0.5
2 -4.7 30 30 120 0.5
5 -5.5 24 24 96 0.5
7 -11.8 30 30 120 0.5
8.1 -13.3 22 22 88 0.5
8.2 -16.3 25 25 100 0.5
8.1 -17.5 22 22 88 0.5
9 -19.6 2.5 2.0 5.7 1.0
10 -21.2 40 40 160 0.5
11 -29.5 16 8 40 1.0
10 -30.0 40 40 160 0.5
Table 4.8: Stiffness parameters inserted in Plaxis. The values presented in red are estimates since no
data was available.
The calculations are performed according to the flowchart presented in Figure 3.4.
46
Results
The final results of the calculations using D-Sheet and Plaxis, using the conventional
design approach, are presented in Table 4.9. In section 4.6 these results are compared
with the outcome of the calculations using the new approach.
Nanchor Mmax Vmax wmax,top wmax,field

[kN] [kNm/m] [kN/m] [mm] [mm]
D-Sheet tangent method 1538 / 1532 1980 477 9 -64
D-Sheet secant method 1562 / 1533 2125 486 15 -70
Plaxis HS 2122 / 1794 2926 587 -91 -195
Table 4.9: Results of calculations applying conventional design approaches. The table gives from left
to right: axial force in both anchors, maximum bending moment and shear force in the wall, and the
maximum wall displacements at the top of the wall and in the field. The displacements w are the horizontal
displacements, where a negative value corresponds with a displacements towards the water.
47
4.5 Design based on alternative design approach using cssm
In this section the design of the quay wall is presented, using the new design method using
critical state soil mechanics theory. Because of the soil profile, consisting of mostly sandy
soil layers, the analysis is performed using purely drained response, and the undrained
analysis using the Shansep theory is not applied here.
Again, calculations have been performed in both D-Sheet and Plaxis.
Based on the analyzed data of the soil investigation of soil layers 5, 9 and 10, a trend
can be observed. For the layers 5 and 11, both dense sand, a φ0cv had been found which is
3-4 degrees less than φ02% , where φ02% stands for the friction angle of the soil at an axial
strain level of 2%. This trend has been used to estimate the critical state friction angles
for the other soil layers consisting of sand, for which no data was available.
For clay, the value found for soil layer 9 is also used for soil layer 11 (both on larger depth
so comparable stress levels.)
With respect to the designs using the conventional design approach, the structural spec-
ifications and load specifications have not changed. The modified strength parameters of
the soil are given in Table 4.10.
Top of layer Material γdry /γsat

Layer φ0 [°] C [kPa] ψ [°]
[m NAP] model [kN/m3 ]
1 +5.0 Sand 17.2 / 19.2 27 0 0
2 -1.5 Sand 17.4 / 19.4 29 0 0
3 -2.7 Sand 17.2 / 19.2 27 0 0
2 -4.7 Sand 17.4 / 19.4 29 0 0
5 -5.5 Sand 17.2 / 19.2 28 0 0
7 -11.8 Sand 17.2 / 19.2 29 0 0
8.1 -13.3 Sand 17.0 / 19.0 24 0 0
8.2 -16.3 Sand 17.2 / 19.2 27 0 0
8.1 -17.5 Sand 17.0 / 19.0 24 0 0
9 -19.6 Clay 15.0 / 15.0 32 0 0
10 -21.2 Sand 18.0 / 20.0 31 0 0
11 -29.5 Clay 17.0 / 17.0 32 0 0
10 -30.0 Sand 18.0 / 20.0 31 0 0
Table 4.10: Normative soil profile and strength parameters based on critical state theory, according to
Rijkswaterstaat (2017).
4.5.1 Design using D-Sheet

Only changes made are the values for the friction angle and the cohesion. Modulus of
subgrade reaction have been held the same as for the conventional design approach, using
a multi-linear stress-displacement diagram.
48
k1 k2 k3
Layer nr.
[kN/m2 /m1 ] [kN/m2 /m1 ] [kN/m2 /m1 ]
1 16000 8000 4000
2 30000 15000 7500
3 16000 8000 4000
2 30000 15000 7500
5 16000 8000 4000
7 20000 10000 5000
8.1 12000 6000 4000
8.2 16000 8000 4000
8.1 12000 6000 3000
9 2000 800 500
10 40000 20000 10000
11 6000 4000 2000
10 40000 20000 10000
Table 4.11: Values of modulus of subgrade reaction in the D-Sheet calculations
4.5.2 Design using Plaxis

Strength parameters of the critical state are used, see Table 4.10. Constant volume of the
soil, soil can still shear but no volume change anymore. Therefore dilatancy angles equal
to zero. Stiffness parameters same as for the conventional design approach, see Table 4.8.
Results
The results of the calculations using both design methods are presented in Table 4.12.
Anchor Forces Mmax Vmax wmax,top wmax,field

D-Sheet sec 1735 / 1608 2695 561 19 -95
Plaxis (HS) 2355 / 1947 3359 658 -124 -243
Table 4.12: Results of calculations applying the new design approach
49
4.6 Results
In this section the outcome of case study 1 is analyzed and discussed.
In the Figures 4.6 up to 4.9 the results of the calculations in terms of forces and displace-
ments of the wall are given, for both the conventional and alternative design approach.
Figure 4.6: Overview of the maximum bending moments in the wall found using the different design
approaches and design methods.
50
Figure 4.7: Overview of the maximum shear force in the wall found using the different design approaches
and design methods.
Figure 4.8: Overview of the anchor forces found using the different design approaches and design methods.
51
Figure 4.9: Overview of the maximum horizontal displacements of the wall, found using the different
design approaches and design methods. Only absolute values are presented.
52
Figure 4.10: Overview of the generated principle axial strains ε1 in case of the new design approach.
The scale on the right varies from 0.0 % (dark blue) till 8.0 % (red).
Interpretation of the results

Based on the results presented in the figures above, two phenomena stand out. First of
all, the results generated using Plaxis gives higher values in terms of both forces and dis-
placements compared to the results generated using D-Sheet. This is especially clear for
the displacements of the wall, which are around 2-3 times as high.
Secondly, the new design approach produces higher values in terms of both sectional forces
and displacements compared to the results of the conventional design approach. This is
the case for both D-Sheet calculations and Plaxis calculations. It can be explained by the
smaller friction angles for the sand layers at the critical state, which results in a decrease
of shear strength mobilization compared to the conventional design approach. The clay
layers can mobilize more shear strength due to higher friction angles, but since there are
only two small clay layers present this results in a minor contribution.
The calculation results of both design methods (see Table 4.9 and Table 4.12) show in-
top
consistent values regarding wmax . This can be explained by the differences in calculation
method regarding the anchor forces and prestressing between D-Sheet Piling and Plaxis.
The relative difference between the two design approaches is presented in Table 4.13.
53
Fanchor,1 Fanchor,2 Mmax Vmax wmax,top wmax,field
D-Sheet 1.27 1.15 1.27 1.36 1.10 1.09
Plaxis 1.15 1.12 1.36 1.25 1.11 1.09
Table 4.13: Relative difference between the outcome of the conventional and alternative design approach,
for both D-Sheet and Plaxis calculations. The ratio is calculated by dividing the value of the alternative
design approach by the value of the conventional design approach.
Possible explanation for the discrepancy in displacements between D-Sheet and Plaxis
results: the assumed stiffness parameters for Plaxis calculation are quite conservative, in
relation to the average values for soils with the same density / cone resistance as presented
in NEN9997-1. Also, the stiffness parameters are determined using data from only 1 or
2 triaxial tests on prepared samples of sand. Using a correlation between stiffness and
density of soil can give more accurate results in terms of displacements of the wall.
4.6.1 Discussion
The new design approach as presented in the WBI2017 is derived specifically for the as-
sessment of slope stability. It is assumed that when sliding of the soil occurs, high strain
levels develop and the critical state of the soil is reached. In this case study however,
there is no sliding plane development of the soil as can be seen in Figure 4.10. The soil is
retained by the wall and can mobilize sufficient shear strength. The maximum principal
axial strain levels reached is 7.3% which means the critical state of the soil is not reached,
even locally. It is questionable if the current application of cssm is a suitable approach for
determining the strength of the soil in case of a retaining wall.
Regarding the critical state friction angle: φ0cv is the end state of the soil, and this should
be independent of the packing (initial state) of the soil. For very loose soil, the critical
state friction angle is likely to be higher than φ02% , and for dense soil the critical state
friction angle is lower than φ02% .
The accuracy of the comparison between the results, and the conclusions following the
comparison, are dependent on the efficiency of the design using the conventional design
approach.
4.6.2 Conclusion
Based on the quantitative results of this case study, the alternative design approach will
produce significantly higher values for both sectional forces and displacements of the quay
wall. This would lead to more conservative design outcomes. Furthermore is the appli-
cation of the alternative design method not suitable in the case of a solely sandy soil
profile.
54
CHAPTER 5
Case study 2 - Amazonehaven modified with normally consolidated clay

layer
This chapter comprises the analysis of a quay wall with a soil profile consisting a normally
consolidated (NC) clay layer. The properties of the clay layer are based on soil inves-
tigation data from the construction project of the quay wall for the Container Transfer
Alblasserdam (CTA). This quay wall is located along the canal Noord between the cities
of Rotterdam and Dordrecht. This area is known to contain sizable layers of soft to very
soft soil.
The properties of the new soil profile used for this case study, and the accompanying
soil parameters are presented in section 5.1. The sections 5.2 and 5.3 give the starting
points for the design calculations and results of both conventional and new design ap-
proach. An analysis of the results and conclusions based on the results of this case study
is presented in section 5.4.2.
55
5.1 Parameters / Boundary conditions
In this case study, the Holocene part of the soil profile consists of sand and soft clay. The
characteristics of the soft clay are based on the soil investigation data obtained during the
construction project Container Terminal Alblasserdam.
5.1.1 Type of soil response

As in compliance with section 2.2.4, an undrained soil response can be assumed if the
hydrodynamic period satisfied the criterion T < 0.01, using Equation 2.5. This criterion
is checked for both the soft clay and the peat. For these soil types, a variety of drainage
lengths and load durations have been taken into account to create an overview which
shows for which conditions an undrained response of the soil can be expected.
For the clay layer the hydrodynamic period is plotted against the duration of the loading.
For varying thickness of the clay layer multiple graphs of the hydrodynamic period T
versus load duration are presented in Figures 5.1 and 5.2. Here the two straight horizontal
lines represent the boundary values for which a clear drained or undrained response can
be expected according to Vermeer and Meier (1998), see also section 2.2.4. For the area
in between these lines such a clear distinction cannot be made and the soil response will
be partially drained, partially undrained.
Figure 5.1: Short-term analysis of the hydrodynamic period (log-scale) versus loading duration for
different thickness of the clay layer
56
Figure 5.2: Long-term analysis of the hydrodynamic period (log-scale) versus loading duration for dif-
ferent thickness of the clay layer
5.1.2 Soil profile

For this case study, a soil profile consisting a normal consolidated clay layer and multiple
sand layers is generated. As shown in the figures above, the clay layer should be modelled
using an undrained analysis for short term loads, and both undrained and drained analysis
for long term loads. The characteristics of this soil profile are presented in Table 5.1.
Top of layer γdry /γsat C k

[m NAP] [kN/m3 ] [kPa] [m/day]
1 +5.0 Sand 17.5 / 19.5 30 0.1 0.86
2 0.0 Clay 15 / 16.5 17.5 12 2.2E-5
3 -18.0 Sand, Pleistocene 18 / 20 32.5 0.1 0.86
Table 5.1: Build-up of the soil profile. For each soil layer, the location and properties are given.
5.1.3 Different scenarios regarding the timing of construction and load-

ing of the wall
After the dredging of the soil in front of the quay wall, the construction phase is finished
and the quay wall can be made ready to use. In case of high demand for quay wall space,
the quay wall will be taken into function as soon as possible, and it is possible that the
design loads of the user phase do occur very soon after construction. In this case the
undrained soil response caused by the dredging of the soil is followed by the undrained
soil response caused by the surcharge behind the quay wall. However, it is also possible
that the quay wall will not be used in the near future due to a lack of demand. Now the
57
soil has time to consolidate and the excess pore pressures will slowly dissipate. To model
both possibilities, the following two scenarios have been used:
• Scenario 1: a fully undrained analysis of the soil response for all construction stages.
• Scenario 2: a drained analysis of the soil response for the construction stages up to
dredging of the soil, and an undrained analysis of the final construction stage where
the design loads are being applied.
58
For the conventional design method a Plaxis model is constructed. The choice which ma-
terial model is used for each soil layer is based on the recommendations given in (Plaxis,
2018). The clay layer are modelled using the Soft Soil material model. The sand layers are
modelled with the hardening soil material model. Since the clay layer shows an undrained
soil response, it is modelled using the undrained (A) function.
The strength and stiffness parameters of the different soil layers that are used in the
calculations are presented in Tables 5.2 and 5.3.
Top of layer Material Drainage γdry /γsat

Layer Soil Type
[m NAP] model Type [kN/m3 ]
1 +5.0 Sand HS Drained 17.5 / 19.5
2 0.0 Clay HS Undrained (A) 15 / 16.5
3 -18.0 Sand HS Drained 18 / 20
Table 5.2: General info of soil profile used in Plaxis
C ref ref POP

Layer φ0 [°] ψ [°] E50 ref Eoed Eur m Rinter
[kPa] [kPa]
1 30 1 0 25 25 100 0.5 0.66 0
2 17.5 12 0 1.6 3.2 12.8 1.0 0.5 0
3 32.5 0.1 2.5 40 40 160 0.5 0.66 0
Table 5.3: Strength and stiffness parameters of the soil layers.
A schematization of the model in Plaxis is given in Figure 6.3. In this figure, the final
situation is depicted, including the acting design loads on the structure.
59
Figure 5.3: Plaxis model of the final construction phase for case study 2. The grey layer represent
moderately dense sand, the light blue layer represents the normally consolidated clay and the orange layer
represents Pleistocene sand.
Results
The final results of both the D-Sheet and Plaxis calculations are presented in Table 5.4.

Plaxis HS scenario 1 2968 / 2702 3685 725 -175 -249
Table 5.4: Results of calculations applying conventional design approaches. The displacements w are the
horizontal displacements, where a negative value corresponds with a displacement towards the water.
60
5.3 Design based on alternative design approach using undrained
analysis combined with cssm
As shown in section 5.1, a soil layer consisting of normally consolidated clay with a drainage
length of 10 meters shows an undrained response for loading situations up to a duration
of multiple years. The undrained behavior of the clay layer is modelled in Plaxis using the
User Defined material model Shansep NGI-ADP. A list of input parameters required for
this model is given in Table 5.5. The model parameters are derived using the correlations
presented in section 3.2.

Normalized stiffness G/su A 70 -
Failure shear strain in compression γf C 10.5 %
Failure shear strain in extension γf E 15.7 %
Failure shear strain in direct simple shear γf DSS 13.1 %
Reference undrained shear strength in active mode su A 108 kN/m2
ref
Reference depth vertref -4 m - NAP
Increase of undrained shear strength over depth su A 4 kN/m2 /m
inc
Ratio between passive and active shear strength su P /su A 0.45 -
Initial mobilized shear strength τ0 /su A 0.5 -
Ratio between shear strength in DSS and active mode su DSS /su A 0.73 -
Poissons ratio ν 0.15 -
Poissons ratio undrained νu 0.495 -
Ratio between su A and σv0 in case of virgin loading (S) alpha 0.27 -
Power. Indicates the dependency of su on OCR m 0.8 -
Minimum value of undrained shear strength su,min 3.0 kN/m2
Table 5.5: Shansep NGI-ADP model parameters and the corresponding values for soft clay in the Rhine
river delta
Symbol Value Unit

Eoed ref 15.0 E3 kN/m2
c0ref 1.0 kN/m2
φ 10.0 °
ψ 0 °
UD-Power 1.0 -
UD-P ref 100 kN/m2
Table 5.6: Shansep NGI-ADP interface parameters of soft clay
Input values extracted from soil investigation data (see Appendix D):
61
Undrained secant stiffness E50 u 22.7 MPa
Original depth of soil sample dsample -7.8 m below surface
Undrained shear strength at depth dsample su A 108 kN/m2
Axial failure strain in compression εf C 10.5 %
Primary compression index Cc 0.053 -
Secondary compression index Cs 0.015 -
Critical state friction angle φ0cv 34 °
Table 5.7: Input used to determine the model parameters of the NGI-ADP Shansep model
Results
In the following table the final results of the design calculations of the alternative design
method are presented. For the two scenarios described in section 5.1.3 the sectional forces
of the wall, the anchor forces and the displacements are given in the following table.

NGI-ADP Shansep Scenario 1 2623 / 2431 2595 608 -137 -189
Table 5.8: Results of calculations using the alternative design approach including the NGI-ADP Shansep
model
62
5.4 Results
In this section the outcome of case study 2 is presented and analyzed. In the Figures 5.4
up to 5.7 the results of the calculations in terms of forces and displacements of the wall
are given, for both the conventional and alternative design approach. Also, the relative
difference between the outcomes of both approaches for each quantity is given in Table
5.9.
Figure 5.4: Overview of the calculated maximum bending moments in the wall for the two design
approaches.
63
Figure 5.5: Overview of the calculated maximum shear force in the wall for the two design approaches.
Figure 5.6: Overview of the calculated anchor forces for the two design approaches.
64
Figure 5.7: Overview of the calculated maximum horizontal displacements of the wall for the two design
approaches.

Scenario 1 0.88 0.90 0.70 0.84 0.78 0.76
Scenario 2 0.87 0.89 0.68 0.75 0.95 0.74
Table 5.9: Relative difference between the outcome of the two design approaches. The ratio is calculated
by dividing the value of the alternative design approach by the value of the conventional design approach.
From the results presented in the figures and table above the following can be observed:
The calculated forces and displacements using the conventional design approach (Plaxis
Hardening soil model, undrained (A)) are larger than the values found using the alterna-
tive design approach (Plaxis NGI-ADP Shansep model). The difference between the two
approaches ranges from 10% for the anchor forces up to 30% for the maximum moment
and field displacement of the wall. This holds for both loading scenarios.
5.4.1 Discussion
For this case study, no calculations using the method of Blum or D-Sheet piling are per-
formed. These methods are not capable to model the undrained behavior of the clay layer
with the same level of accuracy compared to Plaxis and are therefore excluded from the
analysis. Also, there are no triaxial extension test results available for the soil layers of
this case study. The same holds for DSS test results. The ratios between the shear strain
at failure of other clayey soils with similar characteristics like volumetric weight and nor-
65
malized stiffness (Post and Luijendijk, 2018) have been used. This has resulted in a ratio
of 1 : 1.25 : 1.5 for γfC : γfDSS : γfE . The value of γfC is determined using the results of
available laboratory tests.
5.4.2 Conclusion
Based on the quantitative results of this case study, the alternative design approach pro-
duces smaller forces and displacements compared to the conventional design approach. It
could be that the actual stress state of the clay layer is relatively large in this specific
situation. The alternative design method explicitly takes into account the effective stress
levels when calculation the undrained shear strength. However, no direct explanation has
been found for this outcome.
66
CHAPTER 6
Case study 3 - Quay wall situated in overconsolidated clay layer
6.1 Introduction
This chapter comprises the analysis of a quay wall with a soil profile consisting, among
others, an overconsolidated (OC) clay layer. The OC clay is typical for the northern part
of the Netherlands, and is called boulder clay. Here, the Holocene soil layers have been
subjected to large ice loads in the past. For an overview of the soil profile of this case
study, see Figure 3.3. The soil investigation data has been gathered from the construction
project of the new hospital AZG. For this project a deep excavation and construction pit
with retaining walls have been applied (Dijkstra, 2002).
The characteristics of the overconsolidated clay layer, along with an analysis into undrained
soil behavior for different soil types, are presented in section 6.2. In sections 6.3 and 6.4
the starting points for the design calculations plus the results of both the conventional
and new design approach are presented. In section 6.5 the effects of the magnitude of the
load history of the soil is being investigated. Finally, section 6.6 gives an overview of all
the results combined with an analysis and conclusions of this case study.
Relevance of this case

The data of this case study does not originate an actual quay wall design, but from a
building pit in the city center of Groningen. However, in the ports of Delfzijl and also in
the Eemshaven, these quay walls do exist or could be build in the future. Therefore also
a case with this soil type is elaborated.
In Appendix E some lab results are added that form the basis for the derivation of several
model parameters used in the calculations.
67
6.2 Parameters and Boundary conditions
For the purpose of this research, a soil profile is constructed so that an undrained response
of the soil is expected. However, whether a drained or undrained soil response will occur,
depends on several factors such as soil type, permeability and the thickness of the soil
layers. Therefore first the type of soil response for a number of scenarios is analyzed. The
outcome of this is used to construct a soil profile to which an undrained analysis using the
alternative design method can be applied. The soil profile still represents a realistic case
that can occur in reality.
6.2.1 Type of soil response

As in compliance with section 2.2.4, an undrained soil response can be assumed if the
hydrodynamic period satisfied the criterion T < 0.01, using Equation 2.5. This criterion
is checked for both the original clay layers of the Amazonehaven as well as the boulder
clay, where a variety of drainage lengths and load durations have been taken into account.
Soil layer Drainage length [m] Permeability [m/s] Hydrodynamic period

Boulder clay 0.5 1.94E-11 1.67E-2
Boulder clay 1 1.94E-11 4.17E-3
Boulder clay 2 1.94E-11 1.04E-3
Clay (layer 9) 0.3 9.7E-11 1.68E-1
Clay (layer 11) 0.25 8.0E-11 2.65E-1
Sand (layer 5) 3.15 1.0E-5 7.83E+0
Table 6.1: Hydrodynamic period of different soil layers, duration = 3days
For sand layers, the permeability is relatively high and an undrained response of the soil
for typical loading conditions of quay walls is therefore unlikely. This is confirmed by the
value for T of soil layer 5, see Table 6.1. Due to the large drainage length of this layer in
comparison to the other sand layers, it can be seen as representative for all sand layers
present at the location of the Amazonehaven. The sand layers therefore are excluded from
the undrained analyses, and will be modelled using a drained analysis.
For the cohesive soil layers the hydrodynamic period is plotted against the duration of the
loading. A number of scenarios are presented in Figures 6.1 and 6.2. Here the two straight
horizontal lines represent the boundary values for which a clear drained or undrained
response can be expected according to Vermeer and Meier (1998), see also section 2.2.4.
For the area in between these lines such a clear distinction cannot be made and the soil
response will be partially drained, partially undrained.
68
Figure 6.1: Short-term analysis of the hydrodynamic period versus loading duration for different soil
layers. The values of T are presented on logarithmic scale.
From Figure 6.1 it can be seen that the soil layers 9 & 11, which represent existing clay
layers at the location of the Amazonehaven, can be modelled by a drained soil response,
with the exception of very short loading durations up to one week. These layers have a
very small thickness and therefore also very small drainage length, see Table 6.1.
The overconsolidated clay layers however, do show an undrained response for a loading
duration up to several months or years, depending on the thickness of the layer. The
OC clay L=10 graph for example, shows the hydrodynamic period for boulder clay with
a drainage length L=10m, and thus a thickness of the layer of 20 meters. For loading
durations up to 300 days, the soil response is still expected to be undrained. From Figure
6.2 it can be seen that a fully undrained response can be expected for loading durations
till approximately 2 years, and a partially drained partially undrained response for loading
situations up to 50 years, which is the technical lifetime of a quay wall structure.
69
Figure 6.2: Long-term analysis of the hydrodynamic period versus loading duration for different soil
layers. The values of T are presented on logarithmic scale.
6.2.2 Soil profile

For this case study, a soil profile with an OC clay layer of 25 meters is used. Due to the
large thickness of the OC clay layer, it should be modelled using an undrained analysis
for short term loads, and both undrained and drained analysis for long term loads. The
characteristics of this soil profile are presented in Table 6.2.
Top of layer γdry /γsat C k

[m NAP] [kN/m3 ] [kPa] [m/day]
1 +5.0 Boulder clay 17.7 / 17.7 16 34 1.68E-6
2 -20.0 Sand, Pleistocene 18.0 / 20.0 32.5 0.1 0.86
Table 6.2: Soil profile. The properties of the sand layer are the same as those of the Pleistocene sand in
Chapter 4.

In this section the input parameters and results of the conventional design approach are
presented. First a preliminary design calculation performed using the method of Blum
is given. Next, the detailed design calculations are presented. Here a plaxis calculation
is performed where the Hardening Soil material model is used to model the boulder clay.
The option Undrained (A) is used to model the undrained soil behavior of the clay layer
during construction and loading of the quay wall.
70
6.3.1 Preliminary design based on theory of Blum
Parameters
The input parameters used for this case are presented in Table 6.3. These properties are
in accordance with the soil profile given in Table 6.2.
Top of layer Bottom of γdry /γsat

φ0 [°] c0 [kPa] δ [°]
[m NAP] layer [m NAP] [kN/m3 ]
+5.0 -20 17.7 / 17.7 16 34 8
-20 -40 18 / 20 32.5 0 21.7
Table 6.3: Input parameters used to perform Blum calculation
Results
The results of the calculations are presented in Table 6.4.
Variable Value
Anchor Force 2626 kN
Mmax 2065 kNm/m1
Required embedding depth wall 30.0 m - NAP
Table 6.4: Results of Blum calculation
71
For the conventional design method a Plaxis model is constructed where both the boulder
clay and the Pleistocene sand layer have been modelled using the hardening soil material
model. Because of the over-consolidation of the clay layer, the HS material model can in
this case be used for clay as well. (Plaxis, 2018)
However, the clay layer shows an undrained response for the loading situations examined
in this case study. The boulder clay is modelled using the drainage type Undrained (A)
in Plaxis. The strength and stiffness parameters used for the different soil layers are
presented in Tables 6.5 and 6.6.

Layer Soil Type
1 +5.0 Clay, OC HS Undrained (A) 17.7 / 17.7
2 -20.0 Sand HS Drained 18.0 / 20.0
Table 6.5: General info soil profile as used in Plaxis
C ref ref POP

Layer φ0 [°] ψ [°] E50 ref Eoed Eur m Rinter
[kPa] [kPa]
1 16 34 0 8.3 8.3 33.2 1.0 0.5 39
2 32.5 0.1 2.5 29.7 29.7 118.8 0.5 0.66 39
Table 6.6: Strength and stiffness parameters used for the soil layers modelled with the Hardening Soil
model
A schematization of the model in Plaxis is given in Figure 6.3. In this figure, the final
situation is depicted, including the acting design loads on the structure.
72
Figure 6.3: Plaxis model of the final construction phase for case study 3. Blue soil layer is boulder clay,
orange layer is Pleistocene sand.
Results
The final results of the Plaxis calculations are presented in Table 6.7.

Table 6.7: Results of calculations applying conventional design approach. The displacements w are the
horizontal displacements, where a negative value corresponds with a displacement towards the water.
73
6.4 Design based on alternative design approach using undrained
analysis combined with cssm
Design using Plaxis
As shown in section 6.2, a soil layer consisting of boulder clay with a drainage length of
10 meters shows an undrained response for loading situations up to a duration of multiple
years. The undrained behavior of the clay layer below the phreatic surface is modelled in
Plaxis using the User Defined material model Shansep NGI-ADP.

Layer Soil Type
1 +5.0 Clay, OC HS Undrained (A) 17.7 / 17.7
2 0.0 Clay, OC NGI-ADP Shansep Undrained (A) 17.7 / 17.7
3 -20.0 Sand HS Drained 18.0 / 20.0
Table 6.8: Soil profile as specified in Plaxis for the alternative design approach
A list of input parameters used to model soil layer 2 using the NGI-ADP Shansep material
model is given in Table 6.9.
More information about this material model can be found in section 2.5.

Normalized stiffness G/su A 110.1 -
Failure shear strain in compression γf C 3.6 %
Failure shear strain in extension γf E 9.4 %
Failure shear strain in direct simple shear γf DSS 6.5 %
Reference undrained shear strength in active mode su A 115 kN/m2
ref
Reference depth vertref -10.8 m - NAP
Increase of undrained shear strength over depth su A 7 kN/m2 /m
inc
Ratio between passive and active shear strength su P /su A 0.798 -
Initial mobilized shear strength τ0 /su A 0.7 -
Ratio between shear strength in DSS and active mode su DSS /su A 0.90 -
Poissons ratio ν 0.2 -
Poissons ratio undrained νu 0.495 -
Ratio between su A and σv0 in case of virgin loading (S) alpha 0.28 -
Power. Indicates the dependency of su on OCR m 0.8 -
Minimum value of undrained shear strength su,min 1.0 kN/m2
Table 6.9: Shansep NGI-ADP model parameters and the corresponding values for boulder clay
For a number of parameters of the Shansep NGI-ADP material model, correlations have
been determined in the case of insufficient available soil investigation data. See also section
3.2.1. Input values extracted from soil investigation data (see Appendix E):
74
u
Undrained secant stiffness E50 62.4 MPa
Original depth of soil sample dsample -15.8 m below surface
Undrained shear strength at depth dsample su A 189 kN/m2
Axial strain in compression ε1 C 2.4 %
Primary compression index Cc 0.05 -
Secondary compression index Cs 0.02 -
Critical state friction angle φ0cv 16 °
Table 6.10: Input used to determine the parameters of the NGI-ADP Shansep material model
Some notes:
1. There are no triaxial extension test results available for the soil layers of this case
study. The same holds for DSS test results. The ratios between the shear strain at
failure of other clayey soils with similar characteristics like volumetric weight and
normalized stiffness (Post and Luijendijk, 2018) have been used as reference. This
has resulted in a ratio of 1 : 1.8 : 2.6 for γfC : γfDSS : γfE . The value of γfC is
determined using the results of available triaxial compression tests.
Results
In the following table the final results of the design calculations of the alternative design
method are presented. For the two scenarios described in section 5.1.3 the sectional forces
of the wall, the anchor forces and the displacements are given in the following table.

Table 6.11: Results of calculations where the alternative design approach including the NGI-ADP Shansep
model is used
75
6.5 Sensitivity analysis for historic loads
Because the magnitude of historic load and thereby preconsolidation is explicitly taken
into account in the Shansep model, an additional analysis has been performed to look
specifically at the effects of the loading history. Here, for both the conventional and alter-
native design approach calculations have been performed using varying values of the pre-
overburden pressure for the clay layer. The relative difference of both design approaches
is plotted versus the POP values in Figures 6.4 and 6.5.
Figure 6.4: For each quantity the quotient alternative value over conventional value is plotted against
loading history, presented for loading scenario 1.
76
Figure 6.5: For each quantity the quotient alternative value over conventional value is plotted against
loading history, presented for loading scenario 2.
In case of no load history, the alternative design approach will produce the same order
of magnitude for the sectional forces and displacements of the wall, as indicated by a
ratio close to 1. For increasing values of the pre-overburden pressure, the ratio decreases
and the alternative design approach produces lower values than the conventional design
approach.
77
6.6 Results
In this section the outcome of case study 3 is analyzed and discussed.
In the Figures 6.6 up to 6.9 the results of the calculations in terms of forces and displace-
ments of the wall are given, for both the conventional and alternative design approach.
Figure 6.6: Overview of the maximum bending moments in the wall found using the different design
approaches and design methods.
78
Figure 6.7: Overview of the maximum shear force in the wall found using the different design approaches
and design methods.
Figure 6.8: Overview of the anchor forces found using the different design approaches and design methods.
79
Figure 6.9: Overview of the maximum horizontal displacements of the wall, found using the different
design approaches and design methods.
From the results of the figures and table above the following trends can be observed:
• For both loading scenarios, the differences in outcome between the conventional
design approach (Plaxis HS) and the alternative design approach (Plaxis NGI-ADP
Shansep) are small, The differences between the two methods are given in Table
6.12.
• There is a difference between the results of loading scenario 1 and 2. This suggests
that whether drained or undrained behavior of the soil occurs, does influence the soil
stresses and strains that will occur during the lifetime of this quay wall. The results
of scenario 1, where undrained soil behavior is taken into account for the complete
lifetime of the structure, including construction, shows slightly smaller displacements
and smaller forces compared to scenario 2.

Scenario 1 1.06 1.07 0.98 1.03 1.10 1.04
Scenario 2 1.07 1.07 1.00 1.07 1.04 1.00
Table 6.12: Relative difference between the outcome of the two design approaches. The ratio is calculated
by dividing the value of the alternative design approach by the value of the conventional design approach.
6.6.1 Discussion
For this case study, the results of both the method of Blum and D-Sheet are not comparable
with the results of the Plaxis calculations due to the fact that all plaxis calculations do
include at least partially undrained behavior. The results of Blum and D-Sheet do reflect
a fully drained analysis. Also, the assumption that during sliding of the soil very large
strains do occur (for clay 25%) and that the soil is in critical state does not hold for this
80
case study. The quay wall will prevent full sliding of the soil, and the maximum shear
strains reached for this case study are in the order of 8% reached at the passive side of
the quay wall.
For the sensitivity analysis of loading history, only the parameter for POP is changed in
the calculations. In reality, a soil that has experienced a lower or higher maximum pre-
overburden pressure in the past, also possesses different strength and stiffness parameters.
This has not been included in the calculations, which makes the outcome unreliable.
6.6.2 Conclusion
Based on the quantitative results of this case study, the difference in outcome between the
conventional and alternative design approaches is between -2 and 10% for the quantities
displacements and sectional forces. This difference in outcome is not significant when
taking into account the uncertainty that is introduced by 1) the theoretical models used,
2) laboratory results of soil experiments and 3) correlations for model parameters.
81
CHAPTER 7
Discussion
In this chapter the results and approach of the research are discussed. First, in Table 7.1
an overview of the relative difference of the design approaches for each design method is
given.
D-Sheet Plaxis
Case 1 1.10 - 1.36 1.11 - 1.36
Case 2 Not applicable 0.68 - 0.95
Case 3 Not applicable 0.98 - 1.10
Table 7.1: Overview of relative differences of the design approaches for each design method, presenter
for each case study.
The table shows that the ratios of case 1 and case 3 are equal to or larger than 1. This
can be explained due to the fact that the circumstances and boundary conditions of case
1 and also to less extend case 3, do not reflect the conditions for which the alternative
design method originally was developed.
However, case 2 does not fit in this trend. Here the results of the alternative design ap-
proach are lower than those of the conventional design approach. As mentioned before in
section 5.4.2, a direct explanation for this has not been found. Further research in the
direction of local stress states of the soil should be performed to clarify this. Also, the
applicability of the Hardening soil model for soft clay layers such as those in case study 2
can be analysed further.
Based on the results and conclusions of this research, the following remarks and recom-
mendations can be used as input for further research. First, the new design method
incorporates the different locations along a sliding plane, i.e. active shear, direct shear
and passive shear. In case of a quay wall, the slip surface of the soil next to the wall
is either active (on the land side) or passive (on the water side). Only for global slope
failure a full sliding plane is developed and the analogy used in the method for dikes is
representative. However, the method is also used in case of a structural element inside the
dike body to increase stability.
Second, the laboratory tests are not executed according the prescribed standards. Most
82
notable point is the prescribed strain levels for triaxial compression tests and triaxial
extension tests. These prescribed levels have not been reached in the tests, making the
test results not a reliable reflection of the soil properties at critical state. This results in
less reliable model parameters that serve as input for the design calculations. A possible
measure to overcome some of this uncertainty is the use of in-situ measurements of the
displacements of the wall. This data can be used to validate which model better represents
the reality.
Third, the case study with a normal consolidated clay layer indicates that in the absence
of pre-overburden pressure, the alternative design approach produces lower values for both
sectional forces and displacements. The results of case study 3 with an overconsolidated
clay layer show, for a POP value of zero, larger values for the alternative design approach
compared to the conventional design approach. These two outcomes contradict each other.
No direct explanation has been found for this.
83
CHAPTER 8
Conclusion
The objective of this research is to investigate the possibilities to use an alternative design
approach, taking into account anisotrophy and loading history of the soil during undrained
circumstances, for modelling soil behavior in the design processes of a quay wall. Using
the results of the case studies and literature review, the main research question can be
answered.
Can critical state soil mechanics in combination with undrained analysis of

the soil be applied for the design of quay walls, and what are the effects for
the prediction of strength and stability?
Based on the results of the three case studies elaborated in this report, the alternative
design approach applied in this report is not a valid option for a quay wall in sandy soil
due to the absence of undrained soil conditions. For the cases where a clay layer is present,
the hydrodynamic period indicates whether undrained soil behavior will occur. In those
situations the alternative design method could be used.
The new design approach for dikes is developed under the condition that a specific set of
soil investigation data and laboratory tests on the soil samples is being executed. Using
the results of these tests, the model parameters can be determined with more certainty
and the results become more reliable. This approach has to be implemented for quay walls
into further research. Furthermore, the comparability of the failure mechanisms assessed
in this research with respect to those of dikes is limited, due to the absence of a structural
element in dikes.
The results of the case studies show that for a sandy soil profile, the alternative design
approach produces larger values for sectional forces and displacements of the wall than the
conservative design approach. In case of a quay wall situated in a predominantly clayey
soil, the alternative design approach produces results with the same order of magnitude
for sectional forces and displacements of the wall compared to the conventional design
approach. The magnitude of preconsolidation of the soil plays an important role here,
where with increasing values of pre-loading, the results of the alternative design approach
are decreasing in comparison to those of the conventional design approach.
84
Finally, the new assessment approach for dikes is specifically developed for slope stability
of a dike body in case of a flood wave, where excess pore pressures develop inside the
dike body. For retaining structures, the alternative design approach is only valid if an
undrained response of the soil can be expected. This is assessed by determining the hy-
drodynamic period of each soil layer, where the duration of the load acting on the wall
is taken into account. Also, since new design approach is specifically developed for the
assessment of slope stability, the calculations using Plaxis will enable the assessment of
global (slope) stability of the wall, and local shearing and displacements of the soil.
Regarding the assessment of sectional forces of both wall and soil anchors, values for these
quantities are generated by Plaxis which can be used as input for additional calculations
to determine structural safety and stability of the wall.
85
List of Figures
2.1 The left figure shows stress paths of a dilatant soil, the right figure shows
stress paths of a contractant soil. (Verruijt, 2012) . . . . . . . . . . . . . . . 14
2.2 Schematization of Blum for a situation with homogeneous soil profile. In the
left figure the internal stresses in the wall are shown. The right figure shows
the pressures due to soil and groundwater, acting on the wall. (Verruijt,
2012) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Example of a retaining wall in D-Sheet Piling . . . . . . . . . . . . . . . . . 18
2.4 General overview of dike showing the most important elements. Source:
Jonkman et al. (2018) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 Example of a dike with a stability screen. In this case there are grout
anchors applied along the screen, to provide more horizontal support if the
soil body starts shearing. Source: POVM . . . . . . . . . . . . . . . . . . . 21
2.6 Shear strength according to CSSM. (Rijkswaterstaat, 2017) . . . . . . . . . 22
2.7 Hyperbolic stress-strain relation of sand in primary loading for a standard
drained triaxial test. (Plaxis, 2018) . . . . . . . . . . . . . . . . . . . . . . . 24
2.8 Stress-strain relation of clay and peat. The three curves represent the re-
sults of an undrained triaxial compression test (sA u ), an undrained triaxial
extension test (sPu ) and a Direct Simple Shear test (sDSS u ). Gur defines the
isotropic elasticity of the soil. (Naves and Lengkeek, 2017) . . . . . . . . . . 25
3.1 Overview of soil layers and soil types for case study 1. . . . . . . . . . . . . 28
3.4 Overview of construction phases used in the calculations . . . . . . . . . . . 30
3.5 Stress-strain diagram of an CRS test. The blue line represents the test
results. The red line through points A and B represents the tangent line of
the primary loading branch of the diagram. Source: Rijkswaterstaat (2017) 34
4.1 Location of the Amazonehaven in the Port of Rotterdam. Source: Google

Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 Overview Amazonehaven with different quay walls. ’Kade A1’ and ’Kade
A2’ are indicated with the solid green respectively purple line. The dotted
yellow line indicates the old quay wall, which has been demolished after-
wards. Source: Gemeentewerken Rotterdam (2011a) . . . . . . . . . . . . . 37
86
4.3 Cross section of quay wall A1. Source: Gemeentewerken Rotterdam . . . . 39
4.4 Schematization of the variable load behind the quay wall . . . . . . . . . . . 40
4.5 Geotechnical length profile of the Amazonehaven quay wall part A1. Source:
Gemeentewerken Rotterdam . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.6 Overview of the maximum bending moments in the wall found using the
different design approaches and design methods. . . . . . . . . . . . . . . . 50
4.7 Overview of the maximum shear force in the wall found using the different
design approaches and design methods. . . . . . . . . . . . . . . . . . . . . . 51
4.8 Overview of the anchor forces found using the different design approaches
and design methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.9 Overview of the maximum horizontal displacements of the wall, found using
the different design approaches and design methods. Only absolute values
are presented. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.10 Overview of the generated principle axial strains ε1 in case of the new design
approach. The scale on the right varies from 0.0 % (dark blue) till 8.0 %
(red). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.1 Short-term analysis of the hydrodynamic period (log-scale) versus loading

duration for different thickness of the clay layer . . . . . . . . . . . . . . . . 56
5.2 Long-term analysis of the hydrodynamic period (log-scale) versus loading
duration for different thickness of the clay layer . . . . . . . . . . . . . . . . 57
5.3 Plaxis model of the final construction phase for case study 2. The grey
layer represent moderately dense sand, the light blue layer represents the
normally consolidated clay and the orange layer represents Pleistocene sand. 60
5.4 Overview of the calculated maximum bending moments in the wall for the
two design approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.5 Overview of the calculated maximum shear force in the wall for the two
design approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.6 Overview of the calculated anchor forces for the two design approaches. . . 64
5.7 Overview of the calculated maximum horizontal displacements of the wall
for the two design approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.1 Short-term analysis of the hydrodynamic period versus loading duration for
different soil layers. The values of T are presented on logarithmic scale. . . 69
6.2 Long-term analysis of the hydrodynamic period versus loading duration for
different soil layers. The values of T are presented on logarithmic scale. . . 70
6.3 Plaxis model of the final construction phase for case study 3. Blue soil layer
is boulder clay, orange layer is Pleistocene sand. . . . . . . . . . . . . . . . . 73
6.4 For each quantity the quotient alternative value over conventional value is
plotted against loading history, presented for loading scenario 1. . . . . . . 76
6.5 For each quantity the quotient alternative value over conventional value is
plotted against loading history, presented for loading scenario 2. . . . . . . 77
6.6 Overview of the maximum bending moments in the wall found using the
different design approaches and design methods. . . . . . . . . . . . . . . . 78
6.7 Overview of the maximum shear force in the wall found using the different
design approaches and design methods. . . . . . . . . . . . . . . . . . . . . . 79
6.8 Overview of the anchor forces found using the different design approaches
and design methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
87
6.9 Overview of the maximum horizontal displacements of the wall, found using
the different design approaches and design methods. . . . . . . . . . . . . . 80
A.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
A.2 α represents the inclination of the CSL. phi0 is the critical state friction
angle. Source: Appendix F Rijkswaterstaat (2017) . . . . . . . . . . . . . . 94
D.1 Results of lab tests onto soft clay from borehole Container Terminal Al-
blasserdam, part1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
blasserdam, part2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
blasserdam, part3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
blasserdam, part4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
D.5 Results of CU triaxial tests onto soft clay from borehole Container Terminal
Alblasserdam, part1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
D.6 Results of CU triaxial tests onto soft clay from borehole Container Terminal
Alblasserdam, part2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
E.1 Overview of borehole B1,part1 . . . . . . . . . . . . . . . . . . . . . . . . . 106

E.2 Overview of borehole B1,part2 . . . . . . . . . . . . . . . . . . . . . . . . . 107
E.3 Results of triaxial compression test of boulder clay, sample B1S7 . . . . . . 108
E.6 Results of a uniaxial compression test of boulder clay, sample B1S7 . . . . . 111
E.7 Settlement curve of boulder clay, sample B1S7 . . . . . . . . . . . . . . . . 112
E.8 Z - log p curve of boulder clay, sample B1S7 . . . . . . . . . . . . . . . . . . 113
88
List of Tables
2.1 Shansep NGI-ADP model variables. (Brinkgreve and Panagoulias, 2017) . . 25
3.1 Similarities and differences of modelling choices and boundary conditions

for the case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2 Strength and stiffness parameters of the interface of boulder clay for the
Shansep NGI-ADP model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1 Normative soil profile and parameters of quay wall part A1, profile1. In
the 3rd column the volumetric weights of the soil in both dry and saturated
state are given. The 4th and 5th column give the friction angle and cohesion
of the soil for an axial strain level of 2%, based on historical data & triaxial
tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 Structural parameters of quay wall used in Plaxis models . . . . . . . . . . 42
4.3 Structural parameters of grout anchor used in Plaxis models . . . . . . . . . 43
4.4 Input parameters used to perform Blum calculation . . . . . . . . . . . . . 44
4.5 Results of Blum calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.6 Overview of the values for modulus of subgrade reaction using the secant
method. Based on geotechnical report of Amazonehaven quay wall section
A3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.7 Normative soil profile and strength parameters inserted in Plaxis. . . . . . . 46
4.8 Stiffness parameters inserted in Plaxis. The values presented in red are
estimates since no data was available. . . . . . . . . . . . . . . . . . . . . . 46
4.9 Results of calculations applying conventional design approaches. The table
gives from left to right: axial force in both anchors, maximum bending
moment and shear force in the wall, and the maximum wall displacements at
the top of the wall and in the field. The displacements w are the horizontal
displacements, where a negative value corresponds with a displacements
towards the water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.10 Normative soil profile and strength parameters based on critical state the-
ory, according to Rijkswaterstaat (2017). . . . . . . . . . . . . . . . . . . . . 48
4.11 Values of modulus of subgrade reaction in the D-Sheet calculations . . . . . 49
4.12 Results of calculations applying the new design approach . . . . . . . . . . 49
89
4.13 Relative difference between the outcome of the conventional and alternative
design approach, for both D-Sheet and Plaxis calculations. The ratio is
calculated by dividing the value of the alternative design approach by the
value of the conventional design approach. . . . . . . . . . . . . . . . . . . . 54
5.1 Build-up of the soil profile. For each soil layer, the location and properties
are given. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 General info of soil profile used in Plaxis . . . . . . . . . . . . . . . . . . . . 59
5.3 Strength and stiffness parameters of the soil layers. . . . . . . . . . . . . . . 59
5.4 Results of calculations applying conventional design approaches. The dis-
placements w are the horizontal displacements, where a negative value cor-
responds with a displacement towards the water. . . . . . . . . . . . . . . . 60
5.5 Shansep NGI-ADP model parameters and the corresponding values for soft
clay in the Rhine river delta . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.6 Shansep NGI-ADP interface parameters of soft clay . . . . . . . . . . . . . 61
5.7 Input used to determine the model parameters of the NGI-ADP Shansep
model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.8 Results of calculations using the alternative design approach including the
NGI-ADP Shansep model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.9 Relative difference between the outcome of the two design approaches. The
ratio is calculated by dividing the value of the alternative design approach
by the value of the conventional design approach. . . . . . . . . . . . . . . . 65
6.1 Hydrodynamic period of different soil layers, duration = 3days . . . . . . . 68

6.2 Soil profile. The properties of the sand layer are the same as those of the
Pleistocene sand in Chapter 4. . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.3 Input parameters used to perform Blum calculation . . . . . . . . . . . . . 71
6.4 Results of Blum calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.5 General info soil profile as used in Plaxis . . . . . . . . . . . . . . . . . . . . 72
6.6 Strength and stiffness parameters used for the soil layers modelled with the
Hardening Soil model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.7 Results of calculations applying conventional design approach. The dis-
placements w are the horizontal displacements, where a negative value cor-
responds with a displacement towards the water. . . . . . . . . . . . . . . . 73
6.8 Soil profile as specified in Plaxis for the alternative design approach . . . . 74
6.9 Shansep NGI-ADP model parameters and the corresponding values for
boulder clay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.10 Input used to determine the parameters of the NGI-ADP Shansep material
model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.11 Results of calculations where the alternative design approach including the
NGI-ADP Shansep model is used . . . . . . . . . . . . . . . . . . . . . . . . 75
6.12 Relative difference between the outcome of the two design approaches. The
ratio is calculated by dividing the value of the alternative design approach
by the value of the conventional design approach. . . . . . . . . . . . . . . . 80
7.1 Overview of relative differences of the design approaches for each design
method, presenter for each case study. . . . . . . . . . . . . . . . . . . . . . 82
B.1 Correlations to find stiffness parameter for different soil types . . . . . . . . 95
90
B.2 Ratios of stiffness parameters in Plaxis for different soil types. (CUR2003-7,
2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
91
APPENDIX A
Info about current practise CSSM in dikes and embankments
A.1 Approach for determining the critical state friction an-

gle
The shear strength of the soil according to the critical state soil mechanics, is described
using
tmax = s0 sin φ0cv (A.1)
where:
tmax the maximum mobilized shear strength,
1
s0 the mean effective stress (σv0 + σh0 ),
2
φ0cv the critical state friction angle
According to Rijkswaterstaat (2017) the critical state friction angle can be found using a
triaxial test where either:
• the consolidation stress is higher than the yield stress of the soil specimen, or
• the consolidation stress equals the in-situ vertical stress in the soil
The triaxial test has to be performed up to an axial strain level of at least 25% for sand
and clay, and 40% for peat.
In Figure A.1 an example of a stress-strain diagram is given. The diagram gives the
results of an consolidated, undrained triaxial test on a clay specimen of soil layer 9. The
triaxial test is performed up to an axial strain level of 18%. In this example there is a
decreasing trend with increasing axial strain. The value for φcv is therefore probably a
little overestimated if the stress-strain values at 18% axial strain are used.
92
Figure A.1: Example
The result of these test, and other available data are used to determine the critical state
friction angles. For this, Mohr’s Circle is used. Figure A.2 shows the relation between the
critical state friction angle and the angle of the critical state line.
93
Figure A.2: α represents the inclination of the CSL. phi0 is the critical state friction angle. Source:
Appendix F Rijkswaterstaat (2017)
94
APPENDIX B
Finite Element Method
Correlations provided by CUR166 (2008), CUR211 (2014), CUR2003-7 (2003) and Plaxis
(2018) that have been used to determine model parameters for the design calculations:
Stiffness parameters based on triaxial tests

The stiffness parameters are derived using available soil investigation data, where the val-
ues of the secant stiffness E50 and oedometer stiffness Eoed are given. Using the relations
presented in Equations B.1 and B.2, the values for the reference stiffness are determined
(Plaxis, 2018).
c cos φ − σ 0 sin φ m
ref 3
E50 = E50 (B.1)
c cos φ + pref sin φ
σ30
c cos φ − K0nc
sin φ
m
ref
Eoed = Eoed (B.2)
c cos φ + pref sin φ
Stiffness parameters based on correlations

In Tables B.1 and B.2 some empirical formulas are given which can be used to determine
the stiffness parameters, in case triaxial test results are not (sufficiently) available.
ref
Soil type Eoed
50000
Clay (OCR = 1)
Ip
50000
Clay (OCR > 1
Ip
r
pref
Sand (OCR = 1) 3qc
σv0
Table B.1: Correlations to find stiffness parameter for different soil types
95
ref ref ref
Soil type E50 Eoed Eur
Clay (OCR = 1) 2 1 5-8
Clay (OCR > 1) 1 1 4
Sand (OCR = 1) 1 1 4
Table B.2: Ratios of stiffness parameters in Plaxis for different soil types. (CUR2003-7, 2003)
96
APPENDIX C
Additional information case study 1: Amazonehaven
In this appendix additional and more detailed information of the Amazonehaven is given.
This includes more comprehensive descriptions of input values, design calculations and
assumptions made for the calculations.
C.1 D-Sheet calculations

In this section, additional information about the calculations performed in D-Sheet is
given.
D-Sheet options
• Reduce delta friction angles according to CUR: yes
• For the wall friction angle δ, the relationships δ = 2/3 · φ for sand and δ = 1/3 · φ
for clay have been used, according to CUR166 (2008).
• For the vertical balance check, the following parameters are used in the sheet piling
window:
– Height: 1524mm & 1524mm
– Coating area: 2.23 & 1.42 mm2 /mm2
– Section area: 193 & 78 cm2 /m
– Max. point resistance: 0 MPa, see p.61 of PDF Gemeentewerken Rotterdam
(2011a)
– Xi-factor: 0.75
97
APPENDIX D
Additional information case study 2: soft soil consisting of clay and peat
In this appendix additional information and details about the soil investigation data of
case study 2 is given.
D.1 Soil investigation data

In the following figures the results of laboratory tests on soil samples of the reference
project CTA in Alblasserdam are given.
98
Figure D.1: Results of lab tests onto soft clay from borehole Container Terminal Alblasserdam, part1
99
100
101
102
Figure D.5: Results of CU triaxial tests onto soft clay from borehole Container Terminal Alblasserdam,
part1
103
Figure D.6: Results of CU triaxial tests onto soft clay from borehole Container Terminal Alblasserdam,
part2
104
APPENDIX E
Additional information case study 3: Boulder clay
In this appendix additional information and details about the soil investigation data of
case study 3 is given.
E.1 Soil investigation data

In the following figures the results of one borehole of the reference project AZG in Gronin-
gen is given. The properties of the boulder clay are of particular interest for this research.
Also, for one triaxial compression test the lab results that are used for the parameter
determination are presented here.
105
Figure E.1: Overview of borehole B1,part1
106
Figure E.2: Overview of borehole B1,part2
107
Figure E.3: Results of triaxial compression test of boulder clay, sample B1S7
108
109
110
Figure E.6: Results of a uniaxial compression test of boulder clay, sample B1S7
111
Figure E.7: Settlement curve of boulder clay, sample B1S7
112
Figure E.8: Z - log p curve of boulder clay, sample B1S7
113
Bibliography
BAM Infra. Final design Amazonehaven southside quay wall A1. Document code: 2009-
613, 2009.
BAM Infratechniek Midden-Wet bv. Drawing of location CPT’s, boreholes and monitoring
wells. Technical drawing belonging to, 2009.
M. Bolton. The strength and dilatancy of sands. Geotechnique, 36(1):65–78, 1986.
R. Brinkgreve. Selection of soil models and parameters for geotechnical engineering ap-
plication. Soil Constitutive Models: Evaluation, Selection, and Calibration, 1(1):69–98,
2005.
R. Brinkgreve and S. Panagoulias. Shansep NGI-ADP Material Manual. Technical report,
POVM, 2017.
CUR166. ’Handboek Damwandconstructies’. Technical report, SBRCURnet, 2008.
CUR2003-7. Determination Geotechnical Parameters. Technical report, Stichting CUR,
2003.
CUR211. Handbook Quay Walls, Second Edition. Technical report, SBRCURnet, 2014.
Deltares. D-Sheet Piling User Manual. Technical report, Deltares, 2017.
O. Dijkstra. Special solutions for construction AZG Groningen. Educom, 2002.
Fugro GeoServices B.V. Geotechnical investigation broadening Amazonehaven Rotterdam.
Reportnr.: 1010-0204-000, 2011.
Fugro Ingenieursbureau B.V. Drawing of location CPT’s, boreholes and monitoring wells.
Part of 1010-0204-000, 2010.
A. Gaba, S. Hardy, L. Doughty, W. Powrie, and D. Selemetas. Guidance on embedded
retaining wall design. Technical report, CIRIA, 2017.
Gemeentewerken Rotterdam. Final design Amazonehaven southside quay wall A1.
HH2341-0026-RAP, 2011a.
Gemeentewerken Rotterdam. Geotechnical report for Amazonehaven quay wall part A1.
HH2341-0030-RAP, 2011b.
A. L. J. H. Grimstad, G. NGI-ADP: Anisotropic shear strength model for clay. In-
ternational Journal for Numerical and Analytical Methods in Geomechanics, 36(666):
483–497, 2011.
Helpdesk Water. About the legal assessment framework for the primary flood defenses.
Accessed October 23rd , 2017a.
Helpdesk Water. Failure mechanism of macro instability. Accessed October 23rd , 2017b.
J. Jáky. Minimum value of earth pressure. Proc. 2nd Int. Conf. Soil Mech. Found. Eng.,
1948.
114
S. Jonkman, R. Jorissen, T. Schweckendiek, and J. van den Bos. Flood Defenses. Technical
report, Delft University of Technology, 2018.
F. R. Ladd, C.C. New design procedure for stability of soft clays. Journal of the Geotech-
nical Engineering Division, 100(7):763–786, 1974.
E. Meijer. Comparative analysis of design recommendations for Quay Walls. Technical
report, Delft University of Technology, 2006.
N. Mourillon. Stability analysis quay structure at the Amazonehaven, Port of Rotterdam.
Technical report, Delft University of Technology, 2015.
T. Naves and H. Lengkeek. Shansep NGI-ADP validation. Technical report, POVM, 2017.
Plaxis. 2-D Material Models Manual. Technical report, Plaxis bv., 2018.
M. Post and M. Luijendijk. POVM parameter determination FEM. Technical report,
POVM, 2018.
POVM BEEM. POVM calculation tools for FEM. Technical report, Deltares, 2018.
Rijkswaterstaat. Schematiseringshandleiding macrostabiliteit. Technical report, Ministry
of Infrastructure and Water Management, 2017.
T. van Duinen. Back analyses of dikes that withstand a high water level. Technical report,
Deltares, 2013.
T. van Duinen. Undrained shear strength in the assessment of slope stability of water
defenses. Presentation, 2016.
T. van Duinen and H. van Hemert. Stabiliteitsanalyses met ongedraineerde schuifsterkte
voor regionale waterkeringen. Educom, 2013.
P. Vermeer and C. Meier. Stability and deformations in deep excavations in cohesive soils.
Darmstadt Geotechnics, 1(4):177–192, 1998.
A. Verruijt. Soil Mechanics. VSSD, 2012.
D. M. Wood. Soil behaviour and critical state soil mechanics. Cambridge University Press,
2007.
115
116

Thesis Report Final FH Leferink

Uploaded by

Copyright:

Available Formats

Thesis Report Final FH Leferink

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis Report Final FH Leferink

Uploaded by

Copyright:

Available Formats

Delft University of Technology

MSc Thesis report

Research into the application of

4 Case study 1 - Quay wall Amazonehaven 36

5 Case study 2 - Amazonehaven modified with normally consolidated clay

6 Case study 3 - Quay wall situated in overconsolidated clay layer 67

A Info about current practise CSSM in dikes and embankments 92

B Finite Element Method 95

C Additional information case study 1: Amazonehaven 97

E Additional information case study 3: Boulder clay 105

1.1 Problem description & research questions

1.2 Method and scope

2.1 Design of quay walls

Functions of quay walls

Types of quay walls

Sheet pile walls

Structures with relieving platforms

Open berth quays (jetties)

Failure mechanisms of quay walls

2.2.2 Shear strength

τ = c0 + σv0 · tan φ0 (2.1)

2.2.3 Horizontal soil stresses

K0 = (1 − sin φ)OCRsin φ (2.3)

2.2.4 Undrained soil behavior

2.3 Design methods for quay walls

2.3.1 Method of Blum

with: φ0 = effective friction angle of the soil

Modelling undrained soil behavior with the method of Blum

Undrained soil behavior in D-Sheet Piling

2.3.3 FEM - Plaxis

2.4.1 Design process for a dike vs quay wall

Figure 2.6: Shear strength according to CSSM. (Rijkswaterstaat, 2017)

2.5 Material models for describing soil behavior

2.5.2 Hardening Soil

2.5.3 Soft Soil

3.1 Framework case studies

3.1.1 General modelling choices

Software used for the alternative design approach

Figure 3.4: Overview of construction phases used in the calculations

Case study 2 - quay wall Alblasserdam Container Terminal

Case study 3 - quay wall Groningen

3.2.1 General soil parameters

Undrained shear strength

su = qnet /Nkt (3.2)

3.2.2 Shansep model parameters

3.2.3 NGI-ADP model parameters

su P /su A = e−2β (3.7)

Parameter Symbol Value Unit

Case study 1 - Quay wall Amazonehaven

Structural and geometric properties

4.2.1 Soil profile

Model parameters Symbol Value Unit

Table 4.2: Structural parameters of quay wall used in Plaxis models

Table 4.3: Structural parameters of grout anchor used in Plaxis models

Top of layer Bottom of γdry /γsat

Table 4.4: Input parameters used to perform Blum calculation

Some additional factors to be taken into account for the calculation:

Table 4.5: Results of Blum calculation

4.4 Designs based on conventional design approach

Top of layer Material γdry /γsat C