N.Koumparelos Diploma Thesis

National Technical University of Athens Εθνικό Μετσόβιο Πολυτεχνείο
School of Civil Engineering Σχολή Πολιτικών Μηχανικών

Geotechnical Division Τομέας Γεωτεχνικής
Diploma Thesis by
Koumparelos Nikolaos
Supervised by
N. Gerolymos
EFFECTIVE STRESS ANALYSIS OF THE SEISMIC RESPONSE OF

MULTI-BLOCK GRAVITY QUAY WALL: APPLICATION TO
PIRAEUS PORT
ΑΝΑΛΥΣΗ ΤΗΣ ΣΕΙΣΜΙΚΗΣ ΑΠΟΚΡΙΣΗΣ ΛΙΜΕΝΙΚΟΥ

ΚΡΗΠΙΔΟΤΟΙΧΟΥ ΠΟΛΛΑΠΛΩΝ ΤΕΜΑΧΩΝ ΜΕ ΤΗ ΜΕΘΟΔΟ ΤΩΝ
ΕΝΕΡΓΩΝ ΤΑΣΕΩΝ: ΕΦΑΡΜΟΓΗ ΣΤΟ ΛΙΜΑΝΙ TOY ΠΕΙΡΑΙΑ
Διπλωματική Εργασία
Κουμπαρέλου Νικόλαου
Επιβλέπων
Ν. Γερόλυμος
Ιούλιος 2015
Acknowledgements
Upon completing my diploma thesis, I feel the need to thank the people that supported me
during the years of my studying.
This thesis would not have been possible without my supervisor, Assistant Professor N.
Gerolymos, to whom I owe my respectful gratitude not only for the inspiration provided, but
also for his guidance, help and patience. His contribution to solving the challenging obstacles
that arose throughout this thesis was really considerable.
I would also like to express my deepest appreciation and gratitude to Professor G. Gazetas,
for his valuable presence throughout my studies. He has been an endless source of inspiration
and has greatly contributed to the love that I have developed towards Engineering in general
and Geotechnical Engineering in particular.
Finally, above all, I feel the need to thank my family for their consistent support and
encouragement.
1
2
Περιεχόμενα
1. Introduction ............................................................................................................................ 7
1.1 Scope ................................................................................................................................ 7
1.2 Liquefaction ..................................................................................................................... 8
1.3 Interplay between caisson quay wall and soil ................................................................ 15
1.3.1 Historical Incidents .................................................................................................. 15
1.3.2 Interaction between caisson quay wall, seismic earth and water pressures............. 20
2. Numerical Modeling ............................................................................................................ 25
2.1 Finite Element Program ................................................................................................. 25
2.2 Boundaries...................................................................................................................... 36
2.3 Constitutive Model ......................................................................................................... 42
2.3.1 UBCSAND .............................................................................................................. 42
2.3.2 Parameters ............................................................................................................... 46
2.3.3 Undrained behavior ................................................................................................. 47
2.3.4 Rayleigh ................................................................................................................... 49
2.4 Interfaces ........................................................................................................................ 50
2.4.1 Interfaces between the blocks of the quay wall ....................................................... 50
2.4.2 Interfaces at the back and the base of the quay wall ............................................... 50
2.4.3 Interfaces on the frontage of the quay wall ............................................................. 52
3. Quay wall at Piraeus Port ..................................................................................................... 57
3.1 Geometry – Profile ......................................................................................................... 57
3.2 Main Features ................................................................................................................. 60
3.2.1 General – Material set.............................................................................................. 60
3.2.2 Parameters ............................................................................................................... 60
4. Excitations............................................................................................................................ 71
4.1 Input motions.................................................................................................................. 71
4.2 Reduction of excitations ................................................................................................. 73
3
4.3 Polarity ........................................................................................................................... 77
5. Numerical results ................................................................................................................. 81
5.1 Introduction .................................................................................................................... 81
5.2 Fragility criteria .............................................................................................................. 82
5.3 Fragility curves according to PIANC criteria ................................................................ 83
5.3.1 Backfill soil with Dr=40%....................................................................................... 83
5.4 Fragility curves with lower fragility levels .................................................................... 87
5.5 Fragility curves fitting to each type fragility levels ....................................................... 91
6. Conclusion ........................................................................................................................... 97
6.1 General comments .......................................................................................................... 97
6.2 The “0.2g excitation paradox” ....................................................................................... 98
6.3 Verification..................................................................................................................... 98
References .............................................................................................................................. 106
4
EFFECTIVE STRESS ANALYSIS OF THE SEISMIC RESPONSE OF MULTI-BLOCK GRAVITY QUAY WALL: APPLICATION TO PIRAEUS PORT
Chapter 1
Introduction
1. Introduction
1.1 Scope
Experience has shown that port facilities, and especially caisson-type quay walls, are
particularly vulnerable to earthquake related hazards. Motivated by numerous observed cases
where caisson-type quay walls suffered large displacement and rotation during earthquakes,
an illustrative numerical analysis is presented for the response of a typical caisson-type quay
wall section at Piraeus port in Greece.
Utilizing the Byrne's elastoplastic constitutive model, an effective stress dynamic analysis is
performed using as seismic excitation seven recorded motions of the seismic environment
throughout the world. These excitations are applied to three different types of the backfill soil
regarding the relative density (Dr). Primarily, the results emphasize the role of relative
density and time history of excitation on the horizontal displacement and the tilt of a typical
caisson-type quay wall. Supplementary results demonstrate secondary deformations of a
caisson-type quay wall, like the settlement of the backfill soil and the heave of the soil
surface at the sea bottom, in front of the toe of the wall. The excess pore-water pressure,
positive or negative, build-up during shaking is also examined in extent, shedding light on the
liquefied or non-liquefied regions of the soil profile and the ensuing deformation of the
adjacent quay-wall, in sight of interaction.
The numerical modeling is performed using the two-dimensional finite element program
Plaxis2D in order to extract the results that are of our interest in this thesis, either it is the
time histories of some variables or the ultimate values of others.
Concluding, the main purpose of this thesis is to produce fragility functions for the
aforementioned variables for each backfill soil relative density and therefore to investigate
the general agreement of the numerical results of a typical caisson-type quay wall section at
Piraeus port in Greece with those of different case studies throughout the world.
1. Introduction 7
1.2 Liquefaction
Liquefaction is one of the most important, interesting, complex and controversial topics in
geotechnical earthquake engineering. Its devastating effects sprang to the attention of
geotechnical engineers in a tree-month period in 1964 when the Good Friday earthquake
(Mw=9,2) in Alaska was followed by the Niigata earthquake (Ms==7,5) in Japan. Both
earthquakes produced spectacular examples of liquefaction induced damage, including slope
failures, bridge and building foundation failures and flotation of buried structures. In the 50
years since these earthquakes, liquefaction has been studied extensively by hundreds of
researchers around the world. Different procedures and methods of analysis have been
proposed and a prevailing approach has been slow to emerge.
The term liquefaction has historically been used in conjunction with a variety of phenomena
that involve soil deformations caused by monotonic, transient or repeated disturbance of
saturated cohesionless soils under undrained conditions. The generation of excess pore
pressure under undrained loading conditions is a hallmark of all liquefaction phenomena. The
tendency for dry cohesionless soils to densify under both static and cyclic loading is well
known. When cohesionless soils are saturated, however, rapid loading occurs under
undrained conditions, so the tendency for densification causes excess pore pressure to
increase and effective stresses to decrease. Liquefaction phenomena that result from this
process can be divided into two main groups: flow liquefaction and cyclic mobility
Flow liquefaction produces the most dramatic effects of all the liquefaction related
phenomena. It causes tremendous instabilities known as flow failures. Flow liquefaction can
occur when the shear stress required for static equilibrium of a soil mass (the static shear
stress) is greater than the shear strength of the soil in its liquefied state. Once triggered the
large deformations produced by flow liquefaction are actually driven by static shear stresses.
The cyclic stresses may simply bring the soil to an unstable state at witch its strength drops
sufficiently to allow the static stresses to produce the flow failure. Flow liquefaction failures
are characterized by the sudden nature of their origin, the speed with which they develop and
the large distance over which the liquefied materials often move.
Cyclic mobility is another phenomenon that can also produce large permanent deformations
during earthquake shaking. In contrast to flow liquefaction, cyclic mobility occurs when the
static shear stress is less than the shear strength of the liquefied soil. The deformations
produced by cyclic mobility failures develop incrementally during earthquake shaking. In
1. Introduction 8
contrast to flow liquefaction, the deformations produced by cyclic mobility are driven by both
cyclic and static shear stresses. These deformations, termed lateral spreading, can occur on
very gently sloping ground or on virtually flat ground adjacent to bodies of water. When
structures are present, lateral spreading can cause significant damage. A special case of cyclic
mobility is level-ground liquefaction. Because static horizontal shear stresses that could drive
lateral deformations do not exist, level-ground liquefaction can produce large, chaotic
movement known as ground oscillation during earthquake shaking, but produces little
permanent lateral soil movement. Level-ground liquefaction failures are caused by the
upward of water that occurs when seismically induced excess pore pressures dissipate.
Depending on the length of time required to reach hydraulic equilibrium, level-ground
liquefaction failure may occur well after ground shaking has ceased. Excessive vertical
settlement and consequent flooding of low-lying land and the development of sand boils are
characteristics of level-ground liquefaction failure. [Kramer S. L. (1996)]
In other words, liquefaction is a phenomenon caused by the tendency of granular soil to

contract when subjected to monotonic or cyclic shear loading. When this contraction is
prevented or curtailed by the presence of water in the pores, normal stress is transferred from
the soil skeleton to the water. This can cause high excess pore pressures resulting in a very
large reduction in shear stiffness. Large shear strains may occur, and the soil will dilate with
these strains unless the soil is very loose. This dilation causes the porewater pressure to drop
and the stiffness to increase which can limit the strains induced by a load cycle. This behavior
is illustrated in Figures 1.1, 1.2, 1.3 for monotonic loading.
It is this tendency of the soil skeleton to contract and dilate that controls its liquefaction
response. Once the skeleton behavior is modeled, the response under drained, undrained or
coupled stress-flow conditions can be computed by incorporating the bulk stiffness and flow
of the pore fluid. [P. M. Byrne & S. S. Park, M. Beaty]
1. Introduction 9
Figure 1.1: Undrained response of loose sand in simple shear: stress-strain [P. M. Byrne & S. S. Park, M. Beaty]
Figure 1.2: Undrained response of loose sand in simple shear: pore pressure [P. M. Byrne & S. S. Park, M. Beaty]
Figure 1.3: Undrained response of loose sand in simple shear: effective stress response [P. M. Byrne & S. S. Park, M. Beaty]
1. Introduction 10
There different types of stress-strain behavior, illustrated for anisotropically consolidated

specimens in Figure 1.4 were observed. Very loose specimens (such as specimen A in Figure
1.4) exhibited a peak undrained strength at a small shear strain and then collapsed to flow
rapidly to large strains at low effective confining pressure and low large-strain strength.
Dense specimens (specimen B) initially contracted but then dilated until a relatively high
constant effective confining pressure and large-strain strength was reached. At intermediate
densities (specimen C) the exceedance of a peak strength at low strain was followed by a
limited period of strain-softening behavior, which ended with the onset of dilation at
intermediate strains. Further loading produced continued dilation to higher effective
confining pressures and, consequently, higher large-strain strengths. This type of behavior
was termed limited liquefaction. [Kramer S. L. (1996)]
Figure 1.4: Liquefaction, limited liquefaction and dilation in monotonic loading tests [Kramer S. L. (1996)]
1. Introduction 11
Finally, liquefaction phenomena can affect buildings, bridges, buried pipelines and other
constructed facilities in many different ways. Liquefaction can also influence the nature of
ground surface motion. Flow liquefaction can produce massive flow slides and contribute to
the sinking or tilting of heavy structures, the floating of light buried structures and to the
failure or retaining structures. Cyclic mobility can cause slumping of slopes, settlement of
buildings, lateral spreading and retaining wall failure. Substantial ground oscillation, ground
surface settlement, sand boils and post-earthquake stability failure can develop at level-
ground sites.
The occurrence of liquefaction at depth beneath a flat ground surface can decouple the
liquefied soils from the surficial soils and produce large, transient ground oscillations. The
surficial soils are often broken into blocks separated by fissures. These displacements may be
of particular concern for structures supported on pile foundations that extend through
liquefied soils. (Figure 1.5) Sand boils are another liquefaction effect, which is of little
engineering significance by themselves, but they are useful indicators of high excess pore
pressure generation. Porewater draining from the voids of the loose layers can accumulate
beneath the less pervious layers and form water interlayers. Sand boils can develop when the
water interlayers break through the ground surface. (Figure 1.6)
The following pictures depict the aforementioned effects of liquefaction, lateral spreading
(Figure 1.7), ground oscillation and pile foundation damage (Figure 1.8), sand boils (Figure
1.9), settlement of buildings (Figure 1.10).
1. Introduction 12
Figure 1.5: Ground oscillation and potential effect on pile foundation [Kramer S. L. (1996)]
Figure 1.6: Formation of water interlayers [Kramer S. L. (1996)]
1. Introduction 13
Figure 1.7: Collapse of a span of the Nishinomiya Bridge,

Japan [University of Washington]
Figure 1.8: Soil movement about pile, Kobe, Japan [GEER]
Figure 1.9: Sand boil near El Centro, USA, G. Reagor. U.S. Geological Survey [NOAA, NGDC]
Figure 1.10: Building settlement in the 1964 Niigata, Japan [Yale University]
1. Introduction 14
1.3 Interplay between caisson quay wall and soil
1.3.1 Historical Incidents

In past earthquakes, concrete caisson quay walls subjected to strong shaking have repeatedly
suffered substantial outward displacement and rotation, but have only rarely completely
overturned. Nowhere was this more abundantly evident than in the port of Kobe during the
1995 earthquake, where wall displacements reached as much as 5 m with an outward tilting
up to 5o. (Figure 1.11, 1.12)
Liquefaction in the backfill was initially suspected or even proclaimed by some in the
geotechnical community as the main villain behind such large deformations. However,
observations in the field failed to reveal any signs of liquefaction on the backfill surface near
the wall, within a distance of about 2H (i.e. twice the height of the wall)
A more recent case history of a similar example (but smaller in scale) where no liquefaction
occurred next to a substantially displaced quay wall was observed in the Lefkada Ms 6.4
earthquake in Greece (14 August 2003). Having a concrete cross-section of 5 m x 5 m for a
depth of water of only 3.5 m, this wall was supported on sandy gravel and clayey sand. It
displaced laterally about 25 cm (maximum) owing to an acceleration history involving at
least seven cycles of motion with peaks in the range 0.25–0.45g. (Figure 1.13) No sign of
liquefaction was noticed in the immediate backfill, although such evidence did appeared 10–
20 m away from the wall. [Dakoulas, P. & Gazetas, G. (2008)]
The latest incidence, in Greece, took place in Kefalonia at the harbor quay walls of Lixouri,
where a double earthquake (26/1/2014 and 3/2/2014) caused some of the common damages
in a liquefied environment. There were indications of soil liquefaction, namely, sand boil,
lateral spreading and outward displacement of the quay wall. (Figure 1.14)
1. Introduction 15
Figure 1.11: Damaged quay walls and port facilities on Rokko Island, Kobe 1995 [University of Washington]
Figure 1.12: Lateral displacement of a quay wall on Port Island, Kobe 1995 [University of Washington]
1. Introduction 16
Figure 1.13: Lateral spreading at region (a) of the harbor quay walls of Lefkada, Lefkada 2003 [Gazetas G., Dakoulas P.,
Anastasopoulos I.]
Figure 1.14: Damaged quay walls at the port of Lixouri, Kefalonia 2014 [Bouckovalas, G. D.]
1. Introduction 17
Kobe Case History
The numerous case histories from the Kobe port offer a valuable source of field data against
which to check new methods of analysis, and with which to uncover phenomena that are not
well understood at present. Fig. 1.15 shows a plan of the city of Kobe, together with the
approximate location of the strike-slip fault of the 1995 earthquake, and some key points of
interest in this paper.
Figure 1.15: Map of Kobe showing seismologically inferred 1995 earthquake fault, location of studied quay wall (double
arrow), and accelerograph stations showing the recorded peak ground accelerations (PGA) and shaded region of highest
structural damage. The location of the MJMA =7.2 fault is drawn approximately, since it did not emerge on the ground surface
in Kobe; it only did so in Awaji Island (lower left corner). The city of Kobe covers approximately all the area between the
fault and the sea (Osaka Bay) [Dakoulas, P. & Gazetas, G. (2008)]
Most quay walls in Kobe were of the caisson type. They had been designed pseudo-
statically, with seismic coefficients ranging from 0.10 to 0.25, depending on site conditions,
year of construction, and the importance of the facility. They had been placed on top of
gravelly fill consisting of decomposed granite (called locally ‘Masado’), which had
completely replaced the soft clay layer beneath the caisson for improving the bearing
capacity and reducing settlements. The most severe damage occurred in those caisson walls
of Port and Rokko Islands that: (a) were nearly parallel to the coastline (and thus parallel to
the causative fault), and thereby experienced the stronger fault-normal accelerations
(Somerville, 1998); and/or (b) had been designed with a small seismic coefficient, of 0.10 to
0.15. By contrast, the caisson wall of the main wharf at Maya Futo, designed conservatively
with a large seismic coefficient of 0.25 and running almost perpendicular to the fault (and
1. Introduction 18
thereby having been subjected to some less severe accelerations parallel to the fault), did not
experience any visible damage or substantial deformation, remaining operational after the
earthquake. It is worth mentioning that, despite the large deformations, the caissons did not
overturn. Their overall performance can be judged as better than that of the alternative quay
wall system, the anchored sheet-pile wall, which in earlier earthquakes that were much less
devastating than the Kobe 1995 earthquakes were frequently experiencing collapsing failures.
The case history corresponds to the typical quay wall section of Rokko Island, in which both
the foundation and backfill soils are liquefiable. The location of the wall is shown in Fig.
1.15. A cross-section of the quay wall with its deformation recorded after the earthquake is
reproduced from Iai et al. (1998) in Fig. 1.16. The finite-difference discretisation and the
material zones used in our analyses are shown in Fig. 1.17. During the earthquake the wall
top displaced approximately 4 m seaward (exceeding 5 m in a few locations). It settled about
1–2 m and tilted about 4o outwards. Despite these significant movements, the site
investigation showed no collapse of the wall along its entire length. Also, no evidence was
observed of liquefaction either within a zone extending about 30 m behind the wall or near
the toe of the wall in the sea. However, evidence of liquefaction was abundant farther away in
the free field. Investigation by divers cited by Inagaki et al. (1996) revealed substantial
heaving of the foundation rubble at a distance of 2–5 m in front of the toe of the caisson—
indicative of ‘squeezing out’ of the soil underneath the edge (toe) of the tilting caisson.
[Dakoulas, P. & Gazetas, G. (2008)]
Figure 1.16: Cross-section of caisson quay wall RC-5 in Rokko Island and its residual deformation observed after Kobe 1995
earthquake(from Iai et al., 1998) [Dakoulas, P. & Gazetas, G. (2008)]
1. Introduction 19
Figure 1.16: Geometry (in natural scale), finite difference discretisation, and material zones of Rokko Island quay wall
system. Points A, B,C and D and lines aa, bb, cc and dd are for showing details of pore water pressures and displacements
1.3.2 Interaction between caisson quay wall, seismic earth and water pressures
Comprehensive theoretical and experimental research has shed light on the complicated
behavior of the quay wall–soil system. Among the most significant findings of the research
so far are the following.
 The displacement of the wall could be attributed to two factors:

o the significant lateral pressures from the backfill and the large inertia of the
wall itself, as the driving forces
o the strongly inelastic deformation of the foundation soil, allowing the caisson
to move and tilt, as the supporting soil beneath the caisson was ‘pushed out’
(Iai et al., 1998).
 Liquefaction occurred only in the free field away from the wall, not in the backfill
next to the wall. In the foundation soil, on the other hand, very substantial excess
pore water pressures developed, undermining its stiffness and strength. This
facilitated the lateral translation and rotation of the wall, even though ‘complete’
liquefaction did not develop.
 Of the two driving forces, the wall inertia played the most detrimental role, whereas
the earth and water pressures from the backfill were of lesser importance
1. Introduction 20
 The outward movement of the caisson stopped at the end of shaking—an analytical
and experimental finding consistent with the previous conclusions. This provides
further evidence that the whole phenomenon is driven mainly by wall inertia and not
by liquefaction flow. In fact, use of the term ‘lateral spreading’ for the deformations
observed behind the quay walls may not be suitable for describing the whole
phenomenon.
For the quay walls that are of interest here, in particular, the situation becomes far more
complicated because of the following four phenomena that occur simultaneously:
 The development of oscillatory wall inertia loading, which tends to produce outward
displacement and rotation of the wall due to the compliance of the supporting soil.
 The simple-shear deformation of the backfill from the (incident and reflected)
vertically propagating shear waves, which tends to generate positive excess pore
water pressures +Δu in the (usually loose) underwater placed soils. In the free field,
the accumulated build-up of such pressures may lead to liquefaction.
 Extensional normal deformation of the backfill soil adjacent to the wall, as the wall
moves outwards. This tends to generate negative excess pore water pressures -Δu,
which may or may not overshadow the positive seismic pore water pressures,
depending on the amount and speed of the wall movement, as well as on the density
of soil. (Dense dilatant soil in extension may develop exceedingly large negative
water pressures.)
 The tendency to continuous dissipation and redistribution of pore water pressures
(flow in two dimensions), eventually resulting in a detrimental ‘contamination’ of the
regions of negative excess pore water pressure (which therefore become neutralized
or even change their sign to positive) towards the end of shaking.
Evidently, the whole problem is very complex, and with the present state of knowledge there
is no clear understanding of how sensitive each of the above simultaneous phenomena is to
variations in soil characteristics (imperfectly known in reality) such as the relative density
and the coefficient of permeability of the various constituent soils.
1. Introduction 21
1. Introduction 22
Chapter 2
Numerical Modeling
2. Numerical Modeling
2.1 Finite Element Program

PLAXIS 2D is a special purpose two-dimensional finite element program used to perform
deformation and stability analysis for various types of geotechnical applications. Real
situations may be modeled either by a plane strain or an axisymmetric model. The program
uses a convenient graphical user interface that enables the user to quickly generate a
geometry model and finite element mesh based on a representative vertical cross section of
the situation at hand.
The user interface consists of two sub-programs (Input and Output)
The Input program is a pre-processor, which is used to define the problem geometry, to
create the finite element mesh and to define calculation phases.
The Output program is a post-processor, which is used to inspect the results of

calculations in a two dimensional view or in cross sections, and to plot graphs (curves) of
output quantities of selected geometry points.
A general overview for starting the input program, i.e. general model properties and the
layout of the input program is perfectly composed in the [PLAXIS2D Reference Manual].
2. Numerical Modeling 25
General overview of Input program
The soil stratigraphy is defined in the soil mode using the Borehole feature of the program
(Figure 2.1). When a new project is created, the soil contour defined in the Project properties
window is displayed in the draw area. Otherwise, the soil stratigraphy can be defined in the
Structures mode by selecting Create soil polygon option, which creates a general polygon by
specifying the points that define the polygon (Figure 2.2).
In order to define distributed prescribed displacements, the Create line displacement option
should be selected from the menu appearing as the Create prescribed displacement button is
selected. The options for the components of line prescribed displacement are Free, Fixed and
Prescribed. These options can be selected in the Model explorer (Figure 2.2).
Figure 2.1: View of a new project in the Soil mode
Figure 2.2: View of the Structures mode (ground modeling and load modeling)
Soil properties and material properties of structures are stored in material data sets. There are
four different types of material sets grouped as data sets for soil and interfaces, plates,
geogrids, embedded pile rows and anchors (Figure 2.3). The material properties and model
parameters for soil clusters are entered in material data sets. The properties in the data sets
are divided into five tabsheets: General, Parameters, Flow parameters, Interfaces and Initial
(Figure 2.4).
Figure 2.3: Material sets window
Figure 2.4: General and Parameters tabsheet from Soil Window
When the geometry model is fully defined, the geometry has to be divided into finite
elements in order to perform finite element calculations. A composition of finite elements is
called a mesh. The mesh is created in the Mesh mode (Figure 2.5). The mesh should be
sufficiently fine to obtain accurate numerical results. On the other, very fine meshes should
be avoided since this will lead to excessive calculation times. The generation of the mesh is
based on a robust triangulation procedure. The mesh generation process takes into account
the soil stratigraphy as well as all structural objects, loads and boundary conditions.
Figure 2.5: Mesh options window
The Global water level can be used to generate a simple hydrostatic pore pressure distribution
(Phreatic calculation type) for the full geometry (Figure 2.6). The global water level is by
default assigned to all clusters in the geometry.
Figure 2.6: Global water level
Finite element calculations can be divided into several sequential calculation phases. Each
calculation phase corresponds to a particular loading or construction stage. The construction
stages can be defined in the Staged construction mode and the calculation phases are listed in
the Phases explorer.
In the Initial phase (Figure 2.7), all the soil clusters are activated and all the interfaces are
deactivated. It is crucial to mention that in this phase there are soil clusters activated even in
front of the quay walls. In this phase the K0 procedure is used to define the initial stresses for
the model, taking into account the loading history of the soil. The parameters required in the
initial stresses development procedures are defined in the Initial tabsheet of material data sets
for soil and interfaces (Figure 2.8).
Figure 2.7: Initial phase
Figure 2.8: Parameters for the Initial phase
In the first phase (Phase 1) the soil clusters in front of the quay walls are deactivated in order
the water to replace them (Figure 2.9) and all the interfaces remain deactivated. The Plastic
calculation is used to carry out an elastic-plastic deformation analysis in which it is not
necessary to take the change of pore pressure with time into account. An elastic-plastic
deformation analysis where undrained behavior (Undrained (A) or Undrained (B)) is
temporarily ignored, can be defined be selecting the Ignore undrained behavior (A, B)
parameter. In this case the stiffness of water is not taken into account.
Figure 2.9: Fist phase (Phase 1)
Figure 2.10: Parameters for the First phase (Phase 1)
In the second phase (Phase 2) the soil clusters in front of the quay wall remain deactivated
and the interfaces in the back and the base of the quay wall are activated. However, the
interface on the front of the quay wall remains until the ultimate phase deactivated, because
there is not any adjacent soil (Figure 2.11). The Plastic calculation is also in this phase used
and Ignore undrained behavior (A, B) option is also selected (Figure 2.12).
Figure 2.11: Second phase (Phase 2)
Figure 2.12: Parameters for the Second phase (Phase 2)
In the third phase (Phase 3) the soil clusters in front of the quay wall remain deactivated and
the interfaces in the back and the base of the quay wall remain activated. In this last phase the
prescribed line displacement is activated (Figure 2.13). In this phase the Dynamic calculation
is selected in order to consider stress waves and vibrations in the soil. As far as the ignorance
of the undrained behavior is concerned, the Ignore undrained behavior (A, B) option has to
be deselected (Figure 2.14). As for the time step used in a Dynamic calculation, it is constant
and equal to δt = Δt / (m*n), where Δt is the duration of the dynamic loading (Dynamic time
interval), m is the value of Max steps and n is the value of the Number of sub steps parameter.
The Max steps parameter specifies the number of the steps that are stored which can be used
in plots in the Output program. A higher value of the Max steps parameter provides more
detailed plots and animations, however the processing time required by the Output program is
increased as well (Figure 2.15).
Figure 2.13:Third phase (Phase 3)
Figure 2.14: Parameters for the Third phase (Phase 3)
The Model explorer displays information related to the physical entities composing the
model.
In the Dynamics option, the model conditions for a dynamic analysis can be defined at the
extreme boundaries of the model. The options available for the X-axis boundaries are None,
Viscous, Free-field and Tied degrees of freedom. The options available for the Y-axis
boundaries are None, Viscous and Compliant base. (Figure 2.15)
Figure 2.15: Model explorer
The Selection explorer displays information about the selection made in the draw area.
Although the input values of prescribed displacements are specified in the Structures mode,
the activation, deactivation or change of values may be considered in the framework of
Staged construction in the Selection explorer. For a dynamic prescribed displacement
representing an excitation, Displacementx should be Prescribed and Displacementy should be
Fixed (Figure 2.16).
Dynamic multipliers can be assigned to a prescribed displacement. The dynamic multipliers

to be applied in the model can be defined in the DynLineMultiplier subtree under the Line
displacement in the Selection explorer (Figure 2.16). The definition of the multipliers is made
in the Multipliers window. Besides the harmonic signal, there is also the possibility to define
a signal by specifying the values in the table that appears when the corresponding option is
selected in the Signal drop-down menu. Besides defining the signal in the table, there is also
the possibility to read data from a file with a digitized load signal using the Open button in
the toolbar (Figure 2.17).
Figure 2.16: Selection explorer
Figure 2.17: Dynamic multipliers window
After the calculation phases have been defined and before the calculation process is started,
some points may be selected by the user for the generation of load-displacement curves or
stress paths by clicking the Select points for curves button (Figure 2.18).
Figure 2.18: Example of points selected for pre-calculation
Finally, the calculation process can be started by clicking the Calculation button in the tool
bar. As a result the program first performs a check on the ordering and consistency of the
calculation phases. When a calculation phase is selected that has been executed, the tool bar
will show the View results button. Clicking this button will directly display the results of the
selected phase in the Output program (Figure 2.19).
Figure 2.19: Calculation and View results buttons
2.2 Boundaries
With the ultimate purpose to define the boundary conditions that would be selected for the
thesis model, a dynamic calculation has been processed by PLAXIS2D on a soil column
(Figure 2.20), in which two Material models have been used, the linear elastic and the
Hardening Soil small (HSsmall) (non-linear). This layer is based either on a Rigid base or on
a Compliant base. As far as the X-axis is concerned, Free field and Tied degrees of freedom
options have been tested.
The dimensions of the soil column are 2m * 20m, the damping that has been used is ξ = 5%
and the parameters of the soil is presented in Figures 2.21 and 2.22. Regarding the linear
elastic model, the main parameters that are inserted in the Parameters tabsheet, according to
which the others are defined is the velocity Vs=140m/s and the Poisson ratio ν=0, 3. As for
the HSsmall model, the main parameters are E50ref, Eoedref and Eurref, as well as γ0.7.
The fundamental frequency of both models is given by f1 = Vs / (4*H) and the next
frequencies are given by fi = i * f1, i.e. f2 = 2 * Vs / (4*H) etc. The Mode shapes of the first
four frequencies are given in Figures 2.23 and 2.24 for the linear elastic and the Hssmall
model, respectively. This is the first evidence that the dynamic response of the soil column is
in the right direction.
When using the Free field boundary condition we need to create so called node pairs along
the boundary of the model (for this we manually need to create an interface along the
boundary). Between the two nodes of a node pair a viscous damper is created which allows
for transfer of the input and/or free field motion but also allows for absorbing the incoming
waves. Currently however we can by default only see one side of this node pair in the Output
program: this is the side of the viscous damper that also moves due to absorption of
incoming waves. So in general the motion of this node is not equal to the input motion. To be
able to see and to explicitly check our input motion we should make the other node of the
node pair visible in Output. To do so we can make use of a trick with a "dummy plate". This
is why between the X-axis boundaries Free field and Tied degrees of freedom, the later is
selected in this thesis as the simpler one.
Figure 2.20: Soil column
Figure 2.21: Soil parameters for the Linear elastic model
Figure 2.22: Soil parameters for the Hssmall model
f1=1.75 Hz f2=5.24 Hz
f3=8.73 Hz f4=12.22 Hz
Figure 2.23: Mode shapes for the linear elastic model
f1=1.75 Hz f2=5.24 Hz
f3=8.73 Hz f4=12.22 Hz
Figure 2.24: Mode shapes for the HSsmall model
The calculation of the Transfer functions follows in order to prove that the boundaries
conditions used in this thesis are adequately functional. In Figure 2.25 and 2.26 the
amplification functions of the linear elastic and the HSsmall models, respectively, are
presented. In both soil columns Tied degrees of freedom boundary condition is selected for
the X-axis and the Y-axis boundaries are either Rigid base or Compliant base.
In the case where the Free field boundary condition is selected, the soil column has exactly
the same dynamic response.
It can be observed that when compliant base boundary is selected, there is no amplification of
the amplitude of the base, due to the fact that the compliant base absorbs the downward
travelling stress waves so that they are partially reflected, thus part of their energy will be
transmitted through the boundary to continue travelling downward through the base.
Elastic model: rigid base -compliant base

14
Elastic model: rigid base
12 boundary condition
Elastic model: compliant
10 base boundary condition
0
0 2 4 6 8 10 12 14
Figure 2.25: Transfer functions for linear elastic model (rigid base-compliant base)
Hssmall model: rigid base -compliant base

6
Hssmall model: rigid base
boundary condition
5
HSsmall model: compliant
base boundary condition
4
0
0 2 4 6 8 10
Figure 2.26: Transfer functions for HSsmall model (rigid base-compliant base)
In Figure 2.27 Transfer functions for rigid base for both linear elastic and HSsmall model are
compared to show the lower amplification of the HSsmall model, because it is non-linear and
the mitigation of the amplification is obvious especially in the frequencies of resonance.
Rigid base: elastic model-Hssmall model

14
Elastic model: rigid base
12 boundary condition
Hssmall model: rigid base
10
boundary condition
8
0
0 2 4 6 8 10 12 14
Figure 2.27: Transfer functions for rigid base (linear elastic-HSsmall model)
Eventually, in Figure 2.28 there is a comparison between the Amplification functions

calculated by PLAXIS2D and Amplification functions calculated based on the equation:
1
|𝐹2 (𝜔)| = [Kramer S. L. (1996)]
√cos2 𝑘𝐻+sinh2 𝜉𝑘𝐻
It is obvious that there is a perfect agreement between the two calculations.
Consequently, the boundary conditions of the finite element program PLAXIS2D, which are
constantly improved, are of adequate precision. In this thesis, the X-axis boundary Tied
degrees of freedom is selected as the simpler one and the Y-axis boundary Rigid base is
selected because there is a bedrock in the geometry profile of the quay wall at Piraeus Port.
14 Amplification functions (PLAXIS2D - Kramer)

12 Amplification function of linear elastic layer on
rigid base (PLAXIS2D)

comliant base (PLAXIS2D)

rigid base (Kramer)

compliant base (Kramer)
0
0 2 4 6 8 10 12 14
Figure 2.28: Transfer functions for linear elastic model (PLAXIS2D-Kramer)
2.3 Constitutive Model
2.3.1 UBCSAND
For the purpose of this thesis, there has been incorporated into PLAXIS2D an elastic plastic
formulation for the constitutive model UBCSAND in which the yield loci are radial lines of
constant stress ratio and the flow rule is non-associated. This represents a fully coupled
effective stress dynamic analysis procedure for modeling seismic liquefaction.
The simplest realistic model for soil is the classic Mohr-Coulomb elastic-plastic model as
depicted in Figure 2.29. Soils are modeled as elastic below the strength envelope and plastic
on the strength envelope with plastic shear and volumetric strains increments related by the
dilation angle, ψ. This model is really too simple for soils since plastic strains also occur for
stress states below the strength envelope. The UBCSAND stress-strain model described
herein modifies the Mohr-Coulomb model to capture the plastic strains that occur at all stages
of loading. Yield loci are assumed to be radial line of constant stress ratio as shown in Figure
2.30. Unloading is assumed to be elastic. Reloading induces plastic response but with a
stiffened plastic shear modulus.
Figure 2.29: Classic Mohr-Coulomb model [P. M. Byrne & S. S. Park, M. Beaty]
Figure 2.30: UBCSAND model [P. M. Byrne & S. S. Park, M. Beaty]
The plastic shear modulus relates the shear stress and the plastic shear strain and is assumed
to be hyperbolic with stress ratio as shown in Figure 2.31. Moving the yield locus from A to
B in Figure 2.30 requires a plastic shear strain increment, ΔγP, as shown in Figure 2.31, and is
controlled by the plastic shear modulus, GP. The associated plastic volumetric strain
increment, dεvP, is obtained from the dilation angle ψ:
𝛥𝜀𝑣𝑃 = 𝛥𝛾 𝑃 ∗ sin 𝜓 (1)
The dilation angle is based on laboratory data and energy considerations and is approximated
by:
sin 𝜓 = sin 𝜑𝑑 − sin 𝜑𝑐𝑣 (2)
where φcv is the phase transformation or constant volume friction angle and φd describes the
current yield locus. A negative value of ψ corresponds to contraction. Contraction occurs for
stress states below φcv and dilation above as shown in Figure 2.32.
Elastic and plastic properties for the model are defined as follows.
[P. M. Byrne & S. S. Park, M. Beaty]
Elastic Properties
The elastic bulk modulus, B, and shear modulus, Ge, are assumed to be isotropic and stress
level dependent. They are described by the following relations where kB and kG are modulus
numbers, PA is atmospheric pressure, and σ′m is the mean effective stress:
𝜎′𝑚 0.5
𝐵 = 𝑘𝐵 ∗ 𝑃𝐴 ∗ ( ) (3)
𝑃𝐴
𝜎′𝑚 0.5
𝐺 𝑒 = 𝑘𝐺 ∗ 𝑃𝐴 ∗ ( ) (4)
𝑃𝐴
Plastic Properties
The plastic properties used by the model are the peak friction angle φp, the constant volume
friction angle φcv, and plastic shear modulus GP, where
0.5
𝜏
𝐺 𝑃 = 𝐺𝑖𝑃 ∗ (1 − 𝑅𝑓 ) (5)
𝜏𝑓
GiP=α*Ge and α depends on relative density, τ is the current shear stress, τf is the projected
shear stress at failure, and Rf is the failure ratio used to truncate the hyperbolic relationship.
The position of the yield locus φd is known for each element at the start of each time step. If
the stress ratio increases and plastic strain is predicted, then the yield locus for that element is
pushed up by an amount Δφd as given by Equation (6). Unloading of stress ratio is considered
to be elastic. Upon reloading, the yield locus is set to the stress ratio corresponding to the
stress reversal point.
𝐺𝑃
𝛥𝜑𝑑 = ( ) ∗ 𝛥𝛾 𝑃 (6)
𝜎′𝑚
The elastic and plastic parameters are highly dependent on relative density, which must be
considered in any model calibration. These parameters can be selected by calibration to
laboratory test data. The response of the model can also be compared to a considerable
database for triggering of liquefaction under earthquake loading in the field. This database
exists in terms of penetration resistance, typically from cone penetration (CPT) or standard
penetration (SPT) tests. A common relationship between (N1)60 values from the SPT and the
cyclic stress ratio that triggers liquefaction for a magnitude 7.5 earthquake is given by Youd
et al. (2001). Comparing laboratory data based on relative density to field data based on
penetration resistance relies upon an approximate conversion, such as that proposed by
Skempton (1986):
(𝑁1 )60
35 < < 60 (7)
𝐷𝑟2
Model parameters based on penetration resistance and field observation may be useful for
field conditions where it is very difficult to retrieve and test a representative sample.
However, this indirect method is not appropriate for simulation of centrifuge models.
Calibrations for this case should be based on direct laboratory testing of samples that are
prepared in the same manner as the centrifuge model.
2.3.2 Parameters
The PLAXIS2D liquefaction model differentiates the elastic bulk modulus B, and the shear
modulus Ge, [PLAXIS Liquefaction Model]
The elastic behavior which occurs within the yield surface is governed by a non-linear rule.
Two parameters control this non-linear behavior, the elastic bulk modulus K and the elastic
shear modulus G. These two moduli are stress-dependent and the relationships are given in
the following equations (8), (9):
𝑝
𝐾 = 𝐾𝐵𝑒 × 𝑃𝐴 × ( )𝑚𝑒 (8)
𝑃𝑟𝑒𝑓
𝑝
𝐺 = 𝐾𝐺𝑒 × 𝑃𝐴 × ( )𝑛𝑒 (9)
𝑃𝑟𝑒𝑓
where KBe and KGe are the bulk and the shear modulus respectively, at a reference stress level.
The factors ne and me are parameters define the rate of stress dependency of stiffness. In the
literature, the reference stress level pref is commonly taken as the atmospheric pressure
PA=100 kPa.
The hyperbolic hardening rule (Beaty and Byrne, 1998) relates the increment of the sine of
the mobilized friction angle to the plastic shear strain increment as follows (Puebla et al.,
1997):
1
𝛿𝛾 𝑝 = ( ∗ ) × 𝛿 × sin 𝜑𝑚𝑜𝑏 (10)
𝐺
𝑝 𝑝′ sin 𝜑𝑚𝑜𝑏
𝐺 ∗ = 𝐾𝐺 × ( )𝑛𝑝 × {1 − ( ) × 𝑅𝐹 }2 (11)
𝑃𝐴 sin 𝜑𝑝𝑒𝑎𝑘
where KGp is the plastic shear modulus number, np is the plastic shear modulus exponent,
φmob is the mobilized friction angle, which is defined by the stress ratio, φpeak is the peak
friction angle and Rf is the failure ratio nf/nult, ranging from 0.5 to 1.0, where nf is the stress
ratio at failure and nult is the asymptotic stress ratio from the best fit hyperbola.
The input parameters of the UBCSAND are summarized bellow:
 φcv is the constant volume friction angle

 φp is the peak friction angle
 c is the cohesion of the soil
 KBe is the elastic bulk modulus of the soil in a reference level of 100 kPa. It can be
derived from a drained triaxial test with a confining pressure of 100 kPa.
 KGe is the elastic shear modulus of the soil in a reference level of 100 kPa. It can be
related with the KBe using the Poison ratio as shown in Equation:
𝐾𝐵𝑒 2(1 + 𝜈 ′ )
=
𝐾𝐺𝑒 3(1 − 2𝜈 ′ )
 KGp is the plastic shear modulus and has to be extracted after curve fit
 me is the elastic bulk modulus index and has a default value of 0.5
 ne is the elastic shear modulus index and has a default value of 0.5
 np is the plastic shear modulus index and has a default value of 0.5
 Rf is the failure ratio nf/nult
 PA is the atmospheric pressure
 fachard is the densification factor. It is a multiplier that controls the scaling of the
plastic shear modulus during secondary loading. Above 1 the KGp becomes higher and
the behavior stiffer and bellow 1 the KGp becomes lower and the behavior softer
 N160 is the corrected SPT value of the soil.
 facpost is a fitting parameter to adjust post liquefaction behavior
2.3.3 Undrained behavior

The undrained behavior of the soil is treated implicitly by the UBCSAND constitutive model.
Therefore, the increment of the pore water pressure is computed at each step of the analysis.
Considering a saturated soil specimen, the increments in total stress during loading is given
by the following equation:
dp = Ku * dεv (12)
where Ku is the bulk modulus of the undrained soil and dεv the volumetric strain of the soil as
a whole.
The effective stress increment can be computed as follows:
dp' = K’ * dεv (13)
where K’ is the bulk modulus of the soil skeleton and dεv its volumetric strain.
The increments of the pore water pressure are computed with the following equation:
𝐾𝑤
𝑑𝑝𝑤 = × 𝑑𝜀𝜈 (14)
𝑛
where Kw is the bulk modulus of the water, n is the soil porosity and dεv is the volumetric
strain of the fluid.
The relationship between the total stresses, the effective stresses and the pore pressure is
assumed according to Terzaghi's theory (Equation 15). Moreover, the volumetric
compatibility under undrained conditions requires that the equivalent fluid volumetric strain
must be equal to the volumetric strain of the soil skeleton. Equation 16 is finally derived.
dp = dp’ + dpw (15)
𝐾𝑤
= (𝐾𝑢 − 𝐾 ′ ) (16)
𝑛
Once Kw is determined, then the excess pore pressures can be computed in each increment
using Equation 14. The Poisson's ratio for undrained condition is set as ν = 0.495 implicitly
by the model. This value is close to the upper limit (of 0.5) as water is almost incompressible.
Using a value of 0.5 is to be avoided as this is known to cause numerical instabilities. Based
on this Poisson's ratio the bulk modulus of the undrained soil is computed as follows:
2𝐺 𝑒 (1+𝜈𝑢 )
𝐾𝑢 = (17)
3(1−2𝜈𝑢 )
where Ge is the elastic shear modulus.
The drained bulk modulus of the soil skeleton K’ is computed in the same way using the
drained Poisson's ratio which is based on the stress dependent stress moduli (Equation 18).
3𝐾𝑒 −2𝐺 𝑒
𝜈′ = (18)
6𝐾𝑒 +2𝐺 𝑒
2.3.4 Rayleigh
Material damping in dynamic calculations is caused by the viscous properties of soil, friction
and the development of irreversible strains. All plasticity models in PLAXIS2D can generate
irreversible (plastic) strains, and may thus cause material damping. However, the damping is
generally not enough to model the damping characteristics of real soils, especially in cases of
low deformations. For example, most soil models show pure elastic behavior upon unloading
and reloading which does not lead to damping at all. When using these models, the amount of
damping that is obtained depends on the amplitude of the strain cycles and so, these models
do not show material damping. Hence, additional damping is needed to model realistic
damping characteristics of soils in dynamic calculations. This can be done by means of
Rayleigh damping.
The use of additional damping, i.e. Rayleigh α and β, can also lead to numerical convergence.
Rayleigh damping is a numerical feature in which a damping matrix C is composed by
adding a portion of the mass matrix M and a portion of the stiffness matrix K:
C = αM + βK
The parameters α and β are the Rayleigh coefficients and can be specified in the
corresponding cells in the Parameters tabsheet of the Soil window (Figure 2.33).
Figure 2.33: Damping parameters in the General tabsheet
2.4 Interfaces
Interfaces are joint elements to be added to plates or geogrids to allow for a proper modeling
of soil-structure interaction. Interfaces may be used to simulate, for example, the thin zone of
intensely shearing material at the contact between a plate and the surrounding soil. Interfaces
can be created next to plate or geogrid elements or between two soil volumes.
2.4.1 Interfaces between the blocks of the quay wall

The parameters selected for the interface between the blocks of the multi-block quay wall are
shown in the Figure 2.34.
Analyses that have been done within this thesis by PLAXIS2D indicate that the Effective
Young’s modulus should be larger than 100 *103 kN/m2.
Figure 2.34: Interfaces between the blocks of the quay wall
2.4.2 Interfaces at the back and the base of the quay wall
The parameters selected for the interface at the back and the base of the quay wall are shown
in the Figure 2.35 and 2.36.
Analyses that have been done within this thesis by PLAXIS2D indicate that the Effective
Young’s modulus should be larger than 50 *103 kN/m2.
Figure 2.35: Interfaces at the back of the quay wall
Figure 2.36: Interfaces at the base of the quay wall
2.4.3 Interfaces on the frontage of the quay wall

In the thesis model in the first sight, it is obvious that there is not any adjacent soil in the
frontage of the quay wall, thus there is no need for an interface. However, after analyses that
were proceeded within this thesis by PLAXIS2D, it was observed that the lack of interface in
the frontage of the quay wall did not let the blocks to move the one along the other and the
quay wall to move along the base soil. This becomes clear by observing the relative
displacement of the blocks in the case without interface (Figures 2.38, 2.39) and the case with
interface on the frontage (Figures 2.41, 2.42).
In Figures 2.40 and 2.43 it is also observed that the Relative shear stress τrel in the frontage is
eliminated in the case without interface, while in the case with interface the Relative shear
stress τrel in the frontage is finite. This leads to the fact that interface should be used in the
frontage of the quay wall.
Nevertheless, since there is not any adjacent soil, the interfaces can be deactivated when
modeling soil-structure interaction is not desired (Figure 2.37). Nodes generated for
interfaces in mesh generation process are still there. They have stiff elastic behavior and they
are fully permeable.
Figure 2.37: Interface in the frontage of the quay wall
Figure 2.38: Relative displacement of quay wall blocks (without interface)
Figure 2.39: Relative displacement of quay wall blocks (without interface)
Figure 2.40: Relative shear stress τrel in the frontage of the quay wall (without interface)
Figure 2.41: Relative displacement of quay wall blocks (with interface)
Figure 2.42: Relative displacement of quay wall blocks (with interface)
Figure 2.43: Relative shear stress τrel in the frontage of the quay wall (with interface)
Chapter 3
Quay wall at Piraeus Port

3. Quay wall at Piraeus Port
3.1 Geometry – Profile
A typical section of pier II comprising the geometry of the block-type gravity quay wall and
the idealized soil profile is shown in Figures 3.1, 3.2. The distances of the boundaries from
the quay wall are also shown in Figure 3.1.
The layers that constitute the soil profile are the following:
 0m – 20,5m: sandy gravel, φ=35ο, Dr=80%

 20,5m – 25,5m: silty sand, φ=30ο, Dr=60%
 25,5m – 32m: gravelly sand, φ=35ο, Dr=80%
 32m – 37,5m: soft marl, Su=120 kPa
The examined soil profile does not indicate significant liquefaction potential, apart from the
silty sand layer of medium density situated 3m below the base of the quay wall.
Consequently, in order to examine the interaction of the quay wall – soil system and the
potential liquefaction of the soil included, three types of models were dynamically analyzed,
changing only the backfill soil (0m – 20,5m) in each type:
 Model 1: with a significant liquefaction potential, backfill soil with Dr=40%

 Model 2: with a mediocre liquefaction potential, backfill soil with Dr=65%
 Model 3: with a minor liquefaction potential, backfill soil with Dr=80%
The aforementioned Model 1, Model 2 and Model 3 are shown in the Figures 3.3, 3.4 and
3.5, respectively.
3. Quay wall at Piraeus Port 57

Vs (m/s)
9.5 m 70 m 0 250 500
0m
2.5 m
Artificial Fill 0 250 500

Sandy Gravel
17.5 m
φ = 35°
Dr = 80%
25.5 m
17.5 m
20.5 m
Silty sand, φ = 30°, Dr = 60%

25.5 m
Gravelly sand, φ = 35°, Dr = 80%

32 m
Soft marl, Su = 120 kPa
37.5 m
Figure 3.1: Idealized soil profile of pier II of Piraeus Port [Tasiopoulou P., Gerolymos N., Gazetas]
5m
2m
3.5 m 1m
2m 0.5 m
2m 1m
2m
2m 2.5 m
2m 1.5 m
2m
1.5 m
9m
Figure 3.2: Geometry of the block-type gravity quay wall [Tasiopoulou P., Gerolymos N., Gazetas]

Figure 3.3: Idealized soil profile of pier II of Piraeus Port (Model 1)

3.2 Main Features
3.2.1 General – Material set

The first layer of the soil profile has been divided into tree sub-layers, in order to have a
better discretisation of the features of the first layer, i.e. the backfill soil. Therefore, the
ultimate stratigraphy is composed of six layers:
 sandy gravel 1
 sandy gravel 2
 sandy gravel 3
 silty sand
 gravelly sand
 soft marl
The sandy gravel (1,2,3), the silty sand and the gravelly sand Material model has been
selected to be User defined, i.e. UBCSAND model, and their Drainage type has been
Undrained(A). The soft marl Material model is HSsmall and Drainage type Undrained(B).
The concrete of the blocks of the quay wall has as Material model the Linear elastic and as
Drainage type Non-porous.
The interfaces have as Material model the Mohr-Coulomb and as Drainage type Drained.
3.2.2 Parameters
The parameters that have been selected for the layers of the soil profile for each backfill soil
type, i.e. Dr=40%, Dr=65% and Dr=80%, are arrayed in the Figures 3.6 to 3.18
Only the backfill soil layers, i.e. sandy gravel (1,2,3), should alternate their parameters
depending on the backfill soil type, i.e. Dr=40%, Dr=65% and Dr=80%.
Silty sand, gravelly sand and soft marl remain the same at all backfill soil types.
Of course, concrete and interfaces parameters remain as they are.

Dr=40%
Figure 3.6: Parameters of sandy gravel 1 (Dr=40%)

Dr=65%


Dr=80%


Dr=40%, Dr=65% & Dr=80%
Figure 3.15: Parameters of sandy silty sand (Dr=40%, Dr=65% and Dr=80%)
Figure 3.16: Parameters of gravelly sand (Dr=40%, Dr=65% and Dr=80%)

Figure 3.17: Parameters of soft marl (Dr=40%, Dr=65% and Dr=80%)
Figure 3.18: Parameters of concrete (Dr=40%, Dr=65% and Dr=80%)


Chapter 4
Excitation
4. Excitations
4.1 Input motions

The Figures 4.1 to 4.7 show the input motions of the numerical, dynamical analyses.
There have been used seven excitations of real incidents that took place throughout the world,
through the years.
Kalamata(1986), Greece
0,6
0,4
ACCELERATION(g)
0,2
0
-0,2
-0,4
-0,6 0 2 4 6 8 10 12
TIME(s)
Figure 4.1: Excitation of Kalamata(1986), Greece
Kalamata_04g(1986), Greece
0,4
ACCELERATION(g)
0,2
0
-0,2
-0,4
-0,6 0 2 4 6 8 10 12
TIME(s)
Figure 4.2: Excitation of Kalamata_04g(1986), Greece
4. Excitations 71
ACCELERATION(g
0,8
Aegion(1995), Greece
0,6
0,4
0,2
0
-0,2
)
-0,4
-0,6
0 1 2 3 4 5 6
TIME(s)
Figure 4.3: Excitation of Aegion(1995), Greece
Lefkada(2003), Greece
0,6
ACCELERATION(g)
0,4
0,2
0
-0,2
-0,4
-0,6 0 5 10 15 20 25
TIME(s)
Figure 4.4: Excitation of Lefkada(2003), Greece
Erzincan(1992), Turkey
4
ACCELERATION(m/s2)
-2
-4
-6 0 2 4 6 8 10 12 14
TIME(s)
Figure 4.5: Excitation of Erzincan(1992), Turkey
4. Excitations 72
Izmit-Kocaeli(1999), Turkey
0,3
ACCELERATION(g)
0,2
0,1
0
-0,1
-0,2
-0,3 0 5 10 15 20 25 30 35
TIME(s)
Figure 4.6: Excitation of Izmit-Kocaeli(1999), Turkey
Gilroy-Loma Prieta(1989) , Northern California

0,4
0,3
ACCELERATION(g)
0,2
0,1
0
-0,1
-0,2
-0,3
-0,4
0 5 10 15 20 25 30 35 40 45
TIME(s)
Figure 4.7: Excitation of Gilroy-Loma Prieta(1989), Northern California
4.2 Reduction of excitations

The fragility functions were produced based on the maximum acceleration of the excitation
and the scaling of the accelerations was 0,1g – 0,2g – 0,4g – 0,6g. It was vital to make
reduction on scale of the excitations, in order to have in each part of the scale all of the
corresponding excitations with the same maximum acceleration, thus have all the excitations
at the same scale.
This was done by using four reduction-multipliers for each excitation, depending on the real
maximum acceleration of the excitation: The Tables 4.1 – 4.7 show the reduction-multipliers
that make the reduction on scale, accomplished. From these, the -m/s2- reduction-multipliers
4. Excitations 73
have been used within Plaxis2D, due to the fact that Plaxis2D units for excitation
accelerations are m/s2.
The Ux,start,ref option that is used in Plaxis2D (Figure 4.8) is where the reduction-multiplier
of each excitation should be used. Then, the input motion will multiply Ux,start,ref with its
dynamic multipliers (Figure 4.9), thus the scaled excitation will be produced.
Table 4.1: Multipliers used for excitation of Kalamata(1986), Greece within plaxis2D
Kalamata(1986), Greece
Maximum Acceleration=0.39g
g 0.1 0.2 0.4 0.6
multiplier 0.26 0.52 1.03 1.55
m/s2 1 2 4 6
multiplier 2.53 5.06 10.11 15.17
Table 4.2: Multipliers used for excitation of Kalamata_04g(1986), Greece within plaxis2D
Kalamata_04g(1986), Greece
g 0.1 0.2 0.4 0.6
multiplier 0.25 0.50 1.01 1.51
m/s2 1 2 4 6
multiplier 2.46 4.93 9.86 14.79
Table 4.3: Multipliers used for excitation of Aegion(1995), Greece within plaxis2D
Aegion(1995), Greece
g 0.1 0.2 0.4 0.6
multiplier 0.16 0.32 0.64 0.96
m/s2 1 2 4 6
multiplier 1.56 3.13 6.25 9.38
4. Excitations 74
Table 4.4: Multipliers used for excitation of Lefkada(2003), Greece within plaxis2D
Lefkada(2003), Greece
g 0.1 0.2 0.4 0.6
multiplier 0.22 0.44 0.87 1.31
m/s2 1 2 4 6
multiplier 2.14 4.28 8.57 12.85
Table 4.5: Multipliers used for excitation of Erzincan(1992), Turkey within plaxis2D
Erzincan(1992), Turkey
Maximum Acceleration=3.81m/s2
g 0.1 0.2 0.4 0.6
multiplier 0.03 0.05 0.10 0.16
m/s2 1 2 4 6
multiplier 0.26 0.51 1.03 1.54
Table 4.6: Multipliers used for excitation of Izmit-Kocaeli(1999), Turkey within plaxis2D
Izmit-Kocaeli(1999), Turkey
g 0.1 0.2 0.4 0.6
multiplier 0.45 0.91 1.82 2.73
m/s2 1 2 4 6
multiplier 4.46 8.92 17.84 26.75
Table 4.7: Multipliers used for excitation of Gilroy-Loma Prieta(1989), Northern California within plaxis2D
Gilroy, Loma Prieta , Northern California

g 0.1 0.2 0.4 0.6
multiplier 0.27 0.54 1.09 1.63
m/s2 1 2 4 6
multiplier 2.67 5.34 10.68 16.03
4. Excitations 75
Figure 4.8: Example of input multiplier of excitation’s accelerations within Plaxis2D
Figure 4.9: Dynamic multipliers of the input motion within Plaxis2D
4. Excitations 76
4.3 Polarity
The fact that the X-axis boundaries of our model was chosen to be Tied degrees of freedom
[Plaxis2D, Knowledge Base], meant that the model should be symmetrical to the vertical axis
(Figure 4.10). This led to another positive effect. Every input motion that was used had
different polarity for each of the two quay walls of the model. This leads to a totally different
response of the system quay wall-soil against the input excitation [Gazetas G., Garini E.],
thus having a double response to the same input excitation. This is like two excitations were
used for the same quay wall. In other words, counter to the 7 excitations that were really
used, there have been exacted results from 14 input excitations.
Figure 4.10: Symmetrical to the vertical axis model
4. Excitations 77
4. Excitations 78
Chapter 5
Numerical results
5. Numerical results
5.1 Introduction
The ultimate target of this thesis is to produce fragility functions for some meaningful
variables of the liquefaction phenomenon, such as the residual horizontal displacement of the
quay wall, the residual tilt of the caisson quay wall and the settlement of the backfill. In order
to produce the fragility curves there have to be included the acceptable levels of damage, i.e.
the damage criteria.
A remarkable research about the damage criteria of port structures after a seismic incidence
has been developed by PIANC, the World Association for Waterborne Transport
Infrastructure. Among the variety of port structures, PIANC developed damage criteria also
for gravity quay walls. The caisson type quay wall of this thesis falls to the category of
gravity quay walls.
In the sight of a more sensitive aspect and due to the fact that PIANC damage criteria are
slightly non-conservative, there have been produced fragility curves with lower fragility
levels, i.e. more conservative fragility criteria.
Within this thesis and the types of quay wall – soil models that have been included, even the
more conservative fragility criteria cannot present sufficiently the seismic behavior of the
quay walls that their backfill soil type is rather dense, like the soils with Dr=65% and
Dr=80%. Consequently, fragility curves that follow the behavior of each type of quay wall –
backfill soil model have been produced. The fragility levels have been further lowered to the
levels that perfectly represent the behavior of each type of quay wall – backfill soil model.
5. Numerical results 81
5.2 Fragility criteria

PIANC, the World Association for Waterborne Transport Infrastructure, distinguishes four
degrees of damage criteria for gravity quay walls:
 Degree I: “Serviceable”, where the residual horizontal displacement should be

less than 1,5% of the height of the quay wall and the residual tilt should be
less than 3o
 Degree II: “Repairable”, where the residual horizontal displacement should be
1,5%-5% of the height of the quay wall and the residual tilt should be 3o-5o
 Degree III: “Near collapse”, where the residual horizontal displacement should
be 5%-10% of the height of the quay wall and the residual tilt should be 5o-8o
 Degree IV: “Collapse”, where the residual horizontal displacement should be
larger than 10 % of the height of the quay wall and the residual tilt should be
larger than 8o
Attaching the former damage levels to the Piraeus Port quay wall, the fragility criteria
according to PIANC are the following:
 Degree I: the residual horizontal displacement level of exceedance should be

0,25m and the residual tilt level of exceedance should be 3o
 Degree II: the residual horizontal displacement level of exceedance should be
 Degree III: the residual horizontal displacement level of exceedance should be
 Degree IV: the residual horizontal displacement level of exceedance should be
The lower fragility levels selected for the Degrees I, II, III and IV are 0,25m, 0,75m, 1,5m
and 2,5m, respectively, for the residual horizontal displacement and 1o, 2o, 3oand 4o,
respectively, for the residual tilt.
5.3 Fragility curves according to PIANC criteria
5.3.1 Backfill soil with Dr=40%
5.3.1.1 Horizontal displacement
Residual Horizontal displacement - Dr=40%
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
5.3.1.2 Quay wall tilt
Residual Tilt- Dr=40%
100,00%
90,00%
80,00%
70,00% 3o
60,00% 5o
50,00% 8o
40,00% 12o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
5.3.1.3 Backfill soil settlement
Backfill settlement- Dr=40%
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 3o
60,00% 5o
50,00% 8o
40,00% 12o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0,25m
60,00% 1m
50,00% 1,7m
40,00% 3,5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
5.4 Fragility curves with lower fragility levels
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.75m
50,00% 1.5m
40,00% 2.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 1o
60,00% 2o
50,00% 3o
40,00% 4o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.5m
50,00% 0.75m
40,00% 1m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.75m
50,00% 1.5m
40,00% 2.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 1o
60,00% 2o
50,00% 3o
40,00% 4o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.5m
50,00% 0.75m
40,00% 1m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.75m
50,00% 1.5m
40,00% 2.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.5m
50,00% 0.75m
40,00% 1m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
5.5 Fragility curves fitting to each type fragility levels
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.75m
50,00% 1.5m
40,00% 2.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 1o
60,00% 2o
50,00% 3o
40,00% 4o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.5m
50,00% 0.75m
40,00% 1m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.5m
50,00% 1m
40,00% 1.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 1o
60,00% 2o
50,00% 3o
40,00% 4o
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.25m
60,00% 0.35m
50,00% 0.5m
40,00% 0.75m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.1m
60,00% 0.25m
50,00% 0.5m
40,00% 0.75m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
100,00%
90,00%
80,00%
70,00% 0.1m
60,00% 0.2m
50,00% 0.35m
40,00% 0.5m
30,00%
20,00%
10,00%
0,00%
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
Chapter 6
Conclusion
6. Conclusion
6.1 General comments

The most obvious assumption that can be done is that the magnitude of the seismic
excitations, i.e. the maximum acceleration seems to determine, in most of the cases, the size
of the deformations and thus, increasing the amount of excitations that exceed the fragility
level when the excitation’s maximum acceleration varies from 0.1g to 0.6g.
The whole displacements (the residual horizontal displacement of the quay wall, the residual
tilt of the caisson quay wall and the settlement of the backfill) had a decreasing tendency
from the case with backfill soil relative density Dr=40% to Dr=80%. The influence of a
larger relative density to the deformations, in general, is rather positive, thus leading to
limited deformations.
The residual tilting of the quay wall, as a secondary mean of deformation, and due to its
sensitive interaction with the deformations of the adjacent soil, appears to have a noteworthy
behavior, especially in the cases with backfill soil relative density Dr=65% and Dr=80%. In
some of the excitations, the quay wall seems to have the biggest tilting in the excitations with
maximum acceleration 0.2g and 0.4g, while in others it seems to tilt slightly inward,
especially in cases with maximum acceleration 0.4g and 0.6g. For the purposes of the
investigation of this paradox, an analysis with Free field boundary condition has been
processed. For the mitigation of the deformations at the boundaries of our model, a soil
column with the material of the adjacent soil has been formed in both sides. However, this
procedure did not correct the inward tilting of the quay wall. A view of the soil profile of this
process is shown in Figure 6.1.
Figure 6.1: Free field procedure with soil column in the boundaries
6. Conclusion 97
The residual settlement of the backfill soil is in general agreement with the residual
horizontal displacement of the quay wall, i.e. it decreases while relative density Dr increases
and it increases while maximum acceleration increases.
6.2 The “0.2g excitation paradox”

The main paradox that the numerical results indicate is that the biggest residual horizontal
displacement, in the case with backfill soil relative density Dr=65%, seems to appear when
the maximum acceleration of the excitation is 0.2g.
Apparently, the fragility curves for backfill soil with Dr=65% according to PIANC damage
criteria, the one with level of exceedance the 1.7m and the other with 3.5m, appear to have
their peak at the excitations with maximum acceleration 0.2g.
6.3 Verification
The majority of the expected phenomena in the interaction of the system quay wall-soil have
been observed in the numerical analysis by PLAXIS2D.
The lateral spreading and the outward displacement of the quay wall compose a common
response of the system quay wall-soil. In Figure 6.2 a typical response of the quay wall-soil
system is depicted, which appears in all of the models, independently of the relative density
and the magnitude of the excitation. Of course, this appears to all models as a tendency,
setting aside the size of the deformation.
Figure 6.2: Lateral spreading and outward displacement
The outward displacement and tilting of the quay wall have mainly occurred due to the
inelastic deformation of the foundation soil, i.e. the soil in the front of the toe of the quay
wall heaved substantially. This behavior is depicted in Figure 6.3.
6. Conclusion 98
Figure 6.3: Heaving of the soil in front of the toe of the quay wall
As far as the liquefaction is concerned, the regions and the range of liquefaction depends on
the type of backfill soil relative density, the excitation’s maximum acceleration and the
amount of cycles of triggering that the excitation has. Hence, observing the various cases
there are the following assumptions:
 while the relative density increases from a loose to a dense soil for the same
excitation, the range of liquefaction is moderated (Figure 6.4 and 6.5)
 however, when the excitation is really severe the range of liquefaction tends to take
place at the most liquefiable layer of the soil profile (Figure 6.6)
 it is also obvious that liquefaction rarely appears right behind the quay wall, due to
the fact that negative excess pore pressures –Δu are generated, which may or may not
overshadow the positive seismic excess pore pressures (Figure 6.7). Therefore, it is
expected to have the former behavior at excitations with mediocre maximum
acceleration(0.1g)
 while the excitation’s maximum acceleration increases from 0.1g to 0.6g, the range
of liquefaction is expanded backwards and downwards(Figure 6.8)
 the time histories of the pore pressures right behind the quay wall and at the free field
are given in Figure 6.9 and 6.10 for a representative excitation
6. Conclusion 99
Figure 6.4: Range of liquefaction while relative density increases, for the same excitation (Kalamata, 1986 - 0.4g)
Figure 6.5: Range of liquefaction while relative density increases, for the same excitation (Izmit, 1999 - 0.1g)
6. Conclusion 100
Figure 6.6: Range of liquefaction while relative density increases, for the same excitation (Erzincan, 1992 - 0.4g)
Figure 6.7: Range of liquefaction right behind the quay wall, for excitations of 0.1g (Dr=40%)
6. Conclusion 101
Figure 6.8: Range of liquefaction while max acceleration increases, for the same relative density (Lefkada, 2003 – Dr=40%)
6. Conclusion 102
The inertia of the quay wall strongly influences the whole liquefaction phenomenon, but the
excess pore pressure build-up as well.
The liquefaction region for a representative excitation is rather extend (Figure 6.9), and both
points, the one right behind the quay wall and the other at the free field have been liquefied.
There are represented the time histories of excess pore pressures of two points at the same
depth, i.e. with the same initial mean effective stress σ’ο, in order to have the same
denominator at the excess pore pressure ratio ru=Δu/σ’ο. While the excess pore pressure
build-up at free field appears to be steady from a certain time and then, the excess pore
pressure build-up right behind of the quay wall shows ups and downs according to the
movement of the quay wall, outward or inward (Figure 6.10).
Figure 6.9: Liquefaction right behind the quay wall and at free field (Gilroy, 1989 0.2g - Dr=40%)
Figure 6.10: Time histories of excess pore pressures right behind the quay wall and at free field (Gilroy, 1989 0.2g -
Dr=40%)
6. Conclusion 103
Another remarkable observation is that the residual displacements of the backfill soil extend
all the way to the boundaries of the model when the excitation is strong (Figure 6.11), due to
the liquefaction which ranges from the back of the quay wall to the boundary of the model.
Nevertheless, the main displacements take place from the back of the quay wall to the
midwidth of the model, due to the influence of the outward displacement of the quay wall.
Figure 6.11: Mesh of total displacements Ux for strong excitation
6. Conclusion 104
In conclusion, a simple but substantial observation is that the excitations with the most cycles
of triggering and especially with severe oscillations, i.e. with a big period, were those who
led to the biggest residual horizontal displacement of the quay wall. Therefore, the duration
of the excitation had a substantial influence on the response of the quay wall (Figures 6.12
and 6.13).
25
residual horizontal displacement
20 aegion
erzincan
15
gilroy
izmit
10
kalamata_04g
kalamata_1986
5
lefkada_2003
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 g
Figure 6.12: Residual horizontal displacements for each excitation (Dr = 40%)
4,5
residual horizontal displacement
4
3,5
aegion
3
erzincan
2,5 gilroy
2 izmit
1,5 kalamata_04g
1 kalamata_1986
0,5 lefkada_2003
0
0 0,2 0,4 0,6 0,8 g
Figure 6.13: Residual horizontal displacements for each excitation (Dr = 65%)
6. Conclusion 105
References
Bouckovalas, G. D., The Earthquake of Kefalonia, Liquefaction and damages at the port of
Lixouri, Workshop of Hellenic Society of Earthquake Engineering 10/4/2014
Byrne P. M. & Park S.S., Beaty M.: Seismic liquefaction: centrifuge and numerical modeling
Civil and Environmental Engineering, University of Washington:

http://www.ce.washington.edu/~liquefaction/html/quakes/kobe/kobe.html
http://www.ce.washington.edu/~liquefaction/html/where/where1.html
Dakoulas, P. & Gazetas, G. (2008). Geotechnique 58: Insight into seismic earth and water
pressures against caisson quay walls
Gazetas G., Dakoulas P., Anastasopoulos I.: Failure of Harbor Quaywalls in the Lefkada
14-8-2003 Earthquake
Gazetas G., Garini E. (2014): Notes, Sliding Rocking Soil Dynamics
Geotechnical Extreme Events Reconnaissance (GEER), Building Foundations:

http://www.geerassociation.org/GEER_Post%20EQ%20Reports/Kobe_1995/ch3-3.html
Iai, S., Ichii, K., Liu, H. & Morita, T. (1998). Effective stress analysis of port structures. Soils
Found. (Special issue on geotechnical aspects of the January 17, 1995 Hyogoken-Nambu
earthquake), 2, 97–114.
Inagaki, H., Iai, S., Sugano, T., Yamazaki, H. & Inatomi, T. (1996). Performance of caisson
type quay walls at Kobe port. Soils Found. (Special issue on geotechnical aspects of the
January 17, 1995 Hyogoken-Nambu earthquake), 1, 119–136.
Kramer S. L. (1996): Geotechnical Earthquake Engineering
National Oceanic and Atmospheric Administration (NOAA), National Geophysical Data

Center (NGDC), Natural Hazards: http://www.ngdc.noaa.gov/hazardimages/picture/show/163
PIANC, (World Association for Waterborne Transport Infrastructure): Seismic Design

Guidelines for Port Structures
PLAXIS2D, 2014, Scientific Manual
PLAXIS2D, Knowledge Base: http://kb.plaxis.com/tips-and-tricks/use-dynamic-boundary-

conditions
PLAXIS2D, 2014, Reference Manual
PLAXIS Liquefaction Model UBC3D-PLM
Tasiopoulou P., Gerolymos N., Gazetas G.: Seismic effective stress analysis of Gravity block
type quay walls: Application to Piraeus Port
Yale University: http://classes.yale.edu/03-04/anth254a/gallery/gallery_images/geog-

history/fi/00000014.ht

N.Koumparelos Diploma Thesis

Uploaded by

Document Information

Original Description:

Original Title

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

N.Koumparelos Diploma Thesis

Uploaded by

Copyright:

Available Formats

National Technical University of Athens Εθνικό Μετσόβιο Πολυτεχνείο

School of Civil Engineering Σχολή Πολιτικών Μηχανικών

EFFECTIVE STRESS ANALYSIS OF THE SEISMIC RESPONSE OF

ΑΝΑΛΥΣΗ ΤΗΣ ΣΕΙΣΜΙΚΗΣ ΑΠΟΚΡΙΣΗΣ ΛΙΜΕΝΙΚΟΥ

1.1 Scope ................................................................................................................................ 7

1.2 Liquefaction ..................................................................................................................... 8

1.3 Interplay between caisson quay wall and soil ................................................................ 15

1.3.1 Historical Incidents .................................................................................................. 15

2. Numerical Modeling ............................................................................................................ 25

2.1 Finite Element Program ................................................................................................. 25

2.3 Constitutive Model ......................................................................................................... 42

2.3.1 UBCSAND .............................................................................................................. 42

2.3.2 Parameters ............................................................................................................... 46

2.3.3 Undrained behavior ................................................................................................. 47

2.3.4 Rayleigh ................................................................................................................... 49

2.4 Interfaces ........................................................................................................................ 50

2.4.1 Interfaces between the blocks of the quay wall ....................................................... 50

2.4.3 Interfaces on the frontage of the quay wall ............................................................. 52

3. Quay wall at Piraeus Port ..................................................................................................... 57

3.1 Geometry – Profile ......................................................................................................... 57

3.2 Main Features ................................................................................................................. 60

3.2.1 General – Material set.............................................................................................. 60

3.2.2 Parameters ............................................................................................................... 60

4.1 Input motions.................................................................................................................. 71

4.2 Reduction of excitations ................................................................................................. 73

5. Numerical results ................................................................................................................. 81

5.1 Introduction .................................................................................................................... 81

5.2 Fragility criteria .............................................................................................................. 82

5.3 Fragility curves according to PIANC criteria ................................................................ 83

5.3.1 Backfill soil with Dr=40%....................................................................................... 83

5.3.2 Backfill soil with Dr=65%....................................................................................... 84

5.3.3 Backfill soil with Dr=80%....................................................................................... 86

5.4 Fragility curves with lower fragility levels .................................................................... 87

5.4.1 Backfill soil with Dr=40%....................................................................................... 87

5.4.2 Backfill soil with Dr=65%....................................................................................... 88

5.4.3 Backfill soil with Dr=80%....................................................................................... 90

5.5 Fragility curves fitting to each type fragility levels ....................................................... 91

5.5.1 Backfill soil with Dr=40%....................................................................................... 91

5.5.2 Backfill soil with Dr=65%....................................................................................... 92

5.5.3 Backfill soil with Dr=80%....................................................................................... 94

6.1 General comments .......................................................................................................... 97

6.2 The “0.2g excitation paradox” ....................................................................................... 98

References .............................................................................................................................. 106

In other words, liquefaction is a phenomenon caused by the tendency of granular soil to

There different types of stress-strain behavior, illustrated for anisotropically consolidated

Figure 1.6: Formation of water interlayers [Kramer S. L. (1996)]

Figure 1.7: Collapse of a span of the Nishinomiya Bridge,

Figure 1.8: Soil movement about pile, Kobe, Japan [GEER]

1.3 Interplay between caisson quay wall and soil

1.3.1 Historical Incidents

Kobe Case History

 The displacement of the wall could be attributed to two factors:

[Dakoulas, P. & Gazetas, G. (2008)]

2.1 Finite Element Program

The user interface consists of two sub-programs (Input and Output)

The Output program is a post-processor, which is used to inspect the results of

General overview of Input program

Figure 2.1: View of a new project in the Soil mode

Figure 2.3: Material sets window

Figure 2.4: General and Parameters tabsheet from Soil Window

Figure 2.5: Mesh options window

Figure 2.6: Global water level