National Technical University of Athens Eθνικό Μετςόβιο Πολυτεχνείο

School of Civil Engineering Σχολή Πολιτικών Μηχανικών

Geotechnical Division Τομέασ Γεωτεχνικήσ

Diploma Thesis by
Panopoulos Vasileios
Supervised by
N. Gerolymos

DEVELOPMENT OF AN ALGORITHM FOR COMPUTATION AND VISUALIZATION OF

INTERNAL FORCES IN 3D STRUCTURAL ELEMENTS: APPLICATION TO THE NON-LINEAR
LATERAL RESPONSE OF PILES

11.4

11.2

11

10.8

10.6

10.4

10.2

10

9.8
9.8 10 10.2 10.4 10.6 10.8 11 11.2 11.4

ΑΛΓΟΡΙΘΜΟ΢ ΓΙΑ ΣΟΝ ΤΠΟΛΟΓΙ΢ΜΟ ΚΑΙ ΣΗΝ ΑΠΕΙΚΟΝΙ΢Η ΕΝΣΑΣΙΚΩΝ ΜΕΓΕΘΩΝ ΢Ε

ΣΡΙΔΙΑ΢ΣΑΣΑ ΔΟΜΙΚΑ ΢ΣΟΙΧΕΙΑ: ΕΦΑΡΜΟΓΗ ΢ΣΗ ΜΗ-ΓΡΑΜΜΙΚΗ ΕΓΚΑΡ΢ΙΑ
ΑΠΟΚΡΙ΢Η ΠΑ΢΢ΑΛΟ-ΟΜΑΔΑ΢
Διπλωματική Εργαςία
Πανόπουλου Βαςίλειου
Επιβλέπων
Ν. Γερόλυμοσ

Μάϊοσ 2014
Acknowledgements

Upon completing my diploma thesis, I feel the need to thank the people that
supported and encouraged me throughout my work.

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 arised throughout this thesis was really considerable.

I would also like to express my deepest appreciation and gratitude to Professor G.

Gazetas, for his valuable guidance 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.

I would like to thank assistant Professor I. Anastasopoulos for the interest shown and
all the people of the Geotechnical Department and Soil Mechanics laboratory for
their kind embrace as well as my fellow students and friends for the experiences we
shared.

Finally, above all, I feel the need to thank my family for their consistent support and
encouragement.

1
2
Table of Contents
1 Introduction ............................................................................................................................ 6
1.1 Scope ............................................................................................................................... 6
1.2 Use of Pile Foundations ................................................................................................... 7
1.3 Piles under lateral loading ............................................................................................... 8
1.4 Beam on Winkler foundation ........................................................................................ 10
1.5 Beam – Plate hybrid in 3-D soil elements...................................................................... 11
1.6 Embedded pile ............................................................................................................... 12
2 Macroscopic Mohr-Coulomb based approach of the behaviour of circular piles ................ 20
2.1 Introduction ................................................................................................................... 20
2.2 Elements of Macroelement Modeling........................................................................... 20
2.3 Simplified Constitutive Model for RC Pile Section Behaviour ....................................... 21
2.4 Failure Envelope and Plastic Flow Rule for Piles ........................................................... 23
2.4.3 Finite Element Analysis with PLAXIS........................................................................... 25
2.5 Macroelement Modeling ............................................................................................... 26
3 Internal Force Diagrams Computation ................................................................................. 34
3.1 PLAXIS approach ............................................................................................................ 34
3.2 Voronoi decomposition ................................................................................................. 34
3.3 Numerical Integration ................................................................................................... 35
3.4 Slice Method .................................................................................................................. 36
3.5 Interpolation .................................................................................................................. 36
3.6 Volume method ............................................................................................................. 39
3.7 Validation....................................................................................................................... 41
3.8 Result visualization ........................................................................................................ 42
3.9 Stress distribution.......................................................................................................... 43
4 Application to Single Pile ...................................................................................................... 56
4.1 Limit Equilibrium Approach ........................................................................................... 56
4.2 Finite Element Verification ............................................................................................ 58
4.3 Finite Element Modelling .............................................................................................. 59
4.4 Results ........................................................................................................................... 59
5 Application to Pile Group ..................................................................................................... 78
5.1 Limit Equilibrium Approach ........................................................................................... 78
5.2 Finite Element Verification ............................................................................................ 78
APPENDIX ................................................................................................................................ 99

3
Slice method ........................................................................................................................ 99
Volume Method ................................................................................................................ 101

4
Chapter 1

Introduction

5
1 Introduction

1.1 Scope

The purpose of this thesis is to investigate the behaviour of pile foundations under
combined axial, horizontal and moment loading. Assuming undrained conditions,
primary goal is the proposition of a new method, which utilizes the results of the
Finite Element Analysis so as to calculate the structural forces of the pile. This is
attempted in a way so as to overcome the drawbacks of the existing methods and
take fully into consideration the 3-D geometry of the pile, the interaction between
the piles and the interaction between the internal forces. To capture the accurate
pile behaviour, it is used a new macroscopic approach that was developed in
Papakyriakopoulos’ thesis. The validity of this approach is examined in an alternative
scope, so as to both corroborate its results and also extract the pile forces. The
method of the internal forces calculation is implemented in various types of loading,
soil profiles and border conditions. The pile forces of a single pile are examined in
the first stage and, in addition, the interaction of piles in Pile-group is studied.
Emphasis is given to the interaction between the pile axial force and its moment
capacity and moment distribution. Although this method is implemented to piles, its
use can be further extended to any other structural element and provide a tool that
could be proven effective for various applications of civil engineer.

6
1.2 Use of Pile Foundations

Fig 1.1 presents a rough categorization of the types of foundations that are generally
used in order to support structural systems. Surface or shallow embedded
foundations are distinguished by small slenderness ratios, while pile foundations are
generally more slender elements. Caisson foundations lie somewhere in between in
terms of slenderness or embedment; yet their limits are vague. The compressibility
of the soil and the structural element should also be taken into consideration for a
more realistic distinction of different foundation types.

Pile foundations are typically made from steel or reinforced concrete and possibly
timber. They are principally used to transfer the loads from a superstructure,
through weak, compressible strata or water onto stronger, more compact, less
compressible and stiffer soil or rock at depth, increasing the effective size of a
foundation and resisting horizontal loads (Tomlinson & Woodward, 2007). They are
used in very large buildings, off-shore structures, bridge piers and in situations
where the soil under the superstructure is not suitable to prevent excessive
settlement. Piles can be classified by their function:

 End bearing piles are those where most of the friction is developed at the
toe.

 Friction piles are those where most of the pile bearing capacity is developed
by shear stresses along the sides of the pile (Atkinson, 2007).

There are two types of pile foundation installations: driven piles and bored piles:

 Driven piles are normally made from pre-cast concrete which is then
hammered into the ground once on site.

 Bored piles are cast in situ; the soil is bored out of the ground, under reaming
is performed and then the concrete is poured into the hole. Alternatively,
boring of the soil and pouring of the concrete can take place simultaneously,
in which case the piles are called continuous fight augured (CFA) piles.

The choice of pile used depends on the location and type of structure, the ground
conditions, durability of the materials in the environment and cost. Most piles use
some end bearing and some friction, in order to resist the action of loads. Driven
piles are useful in offshore applications, are stable in soft squeezing soils and can
densify loose soil. However, bored piles are more popular in urban areas as there is

7
minimal vibration, they can be used where headroom is limited, there is no risk of
heave and it is easy to vary their length. Deeply embedded foundations have been
consistently used in major offshore structures, where the study of their response
under combined vertical, shear and moment loading is of great importance.

1.3 Piles under lateral loading

In Pile foundations the lateral loads are applied principally in two ways:
 Horizontal static and dynamic loads in the head of the piles, e.g. due to wind,
earthquake, forces from the superstructure, sea waves etc.
 Horizontal loads along the length of the pile-side, e.g. in piled walls, bridge
pier foundations, piles for soil improvement. These piles are usually vertical
and in special circumstances inclined.

The vertical piles undertake horizontal loads with simultaneous bending and lateral
displacement, activating in this way not only their resistance but that of the
surrounding soil too.

The control of the ultimate capacity in horizontal loading must contain:

 The ultimate capacity of the surrounding soil

 The pile resistance as carrying member in bending due to lateral stresses

 The maximum displacement of the pile head, i.e. acceptable from the
superstructure.

The behavior of the piles in horizontal loads depends on many factors as the relative
stiffness of the pile soil system, the stress-strain relation (pile and soil), the soil
resistance and the fixity conditions of the pile head.
The head of the pile, depending on whether the pile is single, belongs to a pile group,
or in other special fixity conditions, might be considered free, pined, or fully fixed.
With respect to the forms of the horizontal loads- displacement diagrams of the total
pile, the piles might be considered as
 Rigid in the case that they rotate around a specific pivot point, without their
significant deformation.

 Flexible in case that their response can be simulated by an elastic beam in

elastic soil.

8
The piles can be categorized also in respect to the ratio L/D (L=length, D=diameter).
A short pile behaves and rotates as a rigid body under lateral loads and has a ratio
L/D<10. When vertical loads are applied, the loads transferred to the tip of the pile
are a percentage of the total. In the case of the long pile (L/D>10), after a certain
length (active length lc) the rest of the pile remains inactive under lateral loading.
Under vertical loading the forces are received by the friction of the pile walls at full
length.

The active Length lc is the minimum length after which the displacement at the pile
head under a certain lateral load remains unaffected.

 ⁄

 ⁄

According to Gazetas (1991) the equations that determine the active length in an
elastic half-space, are dependent of the soil elastic modulus distribution (Fig 1.2.a –
1.2.c).

( )

( )
̃

( )
̃

9
1.4 Beam on Winkler foundation

The Winkler model uses the beam elements to simulate the pile and adopts an
infinite number of closely spaced unconnected springs to model the soil reactions.
Thus the displacement at any point is directly related to the contact pressure at that
point. The stiffness of these springs is uniquely defined by the foundation modulus k.
Later, improved theories have been introduced on refinement of Winkler’s model, by
visualizing various types of interconnections such as shear layers among the Winkler
springs. These theories have been attempted to find an applicable and simple
representation of foundation materials at the contact area. All these models are
mathematically equivalent, but they differ only in definition of the foundation
parameters. The foundation modeling methods, including the usage of formal
expansions, show that the first order approximation corresponds to the
compressibility term of an improved approximation, including the effect of the shear
interactions. Bernoulli-Euler beam theory, Timoshenko beam theory and refined
beam theory (Reddy’s simplified third-order beam theory) are the mainly used beam
theories for bending analysis.

The Winkler model is frequently adopted to describe and simulate the soil behavior.
Its simplicity allows closed-form solutions to be found for various problems.
However, regardless the validity of the soil response, the Winkler model is generally
founded on the method of calibration of the spring coefficients, the dashpot
coefficients and their set-up. Physics of the soil, its constitutive model that describes
the properties of the soil, its plastic flow rule and its interaction with the pile, are not
examined thoroughly. Instead, the proposed approach offers a more technical
solution that is case sensitive and do not apply at all ranges of pile loading.
Furthermore, the beam elements are one dimensional, so they are not able to
reproduce the pile interaction with other elements, owed to its dimensions. Finally,
the beam elements of the pile, may respond well to moment – shear force
combination, but this is not the case with moment - axial force combination. Elastic
theories are not taking into account the interaction between axial force and the pile
bearing capacity, thus neither the beneficial contribution of the compression force
nor the disadvantageous influence of the tension force are considered.

10
1.5 Beam – Plate hybrid in 3-D soil elements

This method proposes a linear viscoelastic pile. It is represented with a series of 3D

Euler-Bernoulli beam elements. The connection of the beam nodes with the
corresponding peripheral soil nodes is established through appropriate kinematic
constraints in order to properly model the pile geometry. In this way, each pile
section behaves as a rigid disk: rotation is allowed on the condition that the disk
remains always perpendicular to the beam axis, but stretching cannot occur. Finally,
full-bonding conditions are assumed at pile-soil interface, which is a simplification of
reality.

However, a number of drawbacks stem when extracting pile capacity and pile
internal forces from this approach. First, the aforementioned method is widely used
in commercially available codes, as the moment-curvature or axial force – strain
constitutive laws work only with beam elements. Furthermore, the interplay
between axial and bending responses is either not taken into account or is captured
in an oversimplified manner, as mentioned above, on beam on Winkler foundation.

In addition, modeling piles with beam elements is a crude approximation of reality

which may lead erroneous results. This is attributed to the one-dimensional nature
of beams (their thickness, or diameter in case of circular piles, is geometrically zero)
that fails to reproduce the lateral capacity of the pile (which is directly proportional
to pile diameter). Moreover, the pile-to-pile interaction in the case of a pile group
depends on the pile spacing ratio s/d. Since diameter (d) is zero, this ratio cannot be
defined and its influence is not taken into account.

Moreover, the effect of soil confining pressure on the increase of the bending
moment capacity of the pile is totally neglected when the latter is modeled with
beam elements. This effect is amplified in lateral loading due to the development of
large passive pressures on the pile opposed to the direction of its movement.

Finally, any benefits that stem from the -up to a point- 3-D functionality of the pile
due to the rigid disks, perpendicular to the beam along the pile length, they have
limited value to the enhanced validity of the results. This is because the Beam - Plate
hybrid response resembles to the Euler - Bernoulli beam, as the cross sections are
perpendicular to the bending line. This assumption is not true and may produce
results that deviate from the accurate solution, especially in dynamic problems.

11
1.6 Embedded pile

An embedded pile consists of beam elements with embedded interface elements to

describe the interaction with the soil at the pile skin and at the pile foot (bearing
capacity). The material parameters of the embedded pile distinguish between the
parameters of the beam and the parameters of the skin resistance and foot
resistance. The beam elements are considered to be linear elastic and are defined by
the same material parameters as a regular beam element. The interaction of the pile
with the soil at the skin of the pile is described by linear elastic behaviour with a
finite strength and is defined by the parameter , the maximum traction allowed
at the skin of the embedded pile and can vary along the pile. An embedded pile is a
pile composed of beam elements that can be placed in arbitrary direction in the sub-
soil and interacts with the sub-soil by means of special interface elements. The
interaction may involve a skin resistance as well as a foot resistance. Although an
embedded pile does not occupy volume, a particular volume around the pile (elastic
zone) is assumed, in which plastic soil behaviour is excluded. The size of this zone is
based on the (equivalent) pile diameter according to the corresponding embedded
pile material data set. This makes the pile almost behave like a volume pile.
However, installation effects of piles are not taken into account and the pile-soil
interaction is modelled at the center rather than at the circumference. The
installation effects of the embedded pile cannot be considered, so this option should
be primarily used for pile types that cause a limited disturbance of the surrounding
soil during installation, such as some types of bored piles, but obviously not driven
piles or soil displacement piles.

An embedded pile consists of beam elements with special interface elements

providing the interaction between the beam and the surrounding soil. The beam
elements are 3-node line elements with six degrees of freedom (ux, uy and uz) and
three rotational degrees of freedom (φx, φy and φz). Element stiffness matrices are
integrated from the four Gaussian points. The element allows for beam deflections
sue to shearing as well as bending. In addition, the element can change length when
an axial force is applied.

The special interface elements are different from the regular interface elements as
used along walls or volume piles. Therefore, at the position of the beam element
nodes, virtual nodes are created in the soil volume element from the element shape
functions. The special interface forms a connection between the beam element
nodes and these virtual nodes, and thus with all nodes of the soil volume element.
Pile forces are evaluated at the beam element integration points and extrapolated to
the beam element nodes.

12
As described about the aforementioned beam, the case with the embedded pile is
that, regardless the satisfying interaction with the adjacent soil due to the unique
interface elements, its geometry and its constitutive model cannot describe well the
physics and the actual response of the three-dimensional pile. Specifically, the one-
dimensional beam is not able to take into consideration the real interaction between
the soil and the pile, because the pile dimension is neglected

13
14
Figures

15
Figure 1.1. A rough categorization of different foundation types based on their
slenderness or embedment ratio D/B. From the left to the right we can distinguish (a)
piles, (b) deeply embedded foundations and (c) shallow foundations. (after
Gerolymos & Gazetas, 2006)

Figure 1.2. Different distributions of the Elastic Modulus with the depth.

16
 P

Figure 1.3. Beam on Winkler foundation

17
Figure 1.4. Beam – Plate hybrid in 3-D soil elements.

RS

TS;max

RN

KF

Fbot;max

Figure 1.5. Embedded pile.

18
 Chapter 2
Macroscopic Mohr-Coulomb based approach of
the behaviour of circular piles

19
2 Macroscopic Mohr-Coulomb based approach of the behaviour
of circular piles

2.1 Introduction

The macroelement can be thought of as an advanced finite element, and more

precisely a generalized "spring", in which the response of the foundation is described
in terms of generalized force versus generalized displacement. This element is placed
at the base of the superstructure and aims at reproducing nonlinear interaction
phenomena arising at soil – foundation interface. Evidently, the translational and
rotational degrees of freedom of a macro-element are all fully coupled. The concept
of macroelement is not new in geotechnical engineering. It was originally introduced
by Roscoe and Schofield (1956) for shallow foundations. To the best of authors'
knowledge, only recently the idea has been extended to pile foundations (Correia et
al., 2012). In this paper a mathematical framework for macroelement modeling of
single piles is briefly presented, emphasizing the nonlinear behaviour of both the soil
and the pile. The calibration of the model parameters is achieved through
comparisons with 3D finite element analyses with the use of code PLAXIS. Given that
the ultimate lateral capacity of a pile is directly related to its diameter and bending
moment resistance, which in turn is a function of the axial force imposed (or
developed) on the pile, a simple uniaxial stress-strain model based on the Mohr -
Coulomb yield criterion is also developed capable of reproducing the cross-sectional
behaviour of circular reinforced concrete piles in terms of bending moment ─
curvature relationship and bending moment ─ axial force failure envelopes.

2.2 Elements of Macroelement Modeling

Recent research has shown that the use of laws and equations provided by the
theory of elastoplasticity can be directly applicable to the analysis of foundations in
cohesive soil under undrained loading conditions (Martin and Houlsby, 2000). It has
been demonstrated that this approach provides better results in comparison to
Winkler based model, as it is capable of realistically representing the coupling
between the various degrees of freedom. Within the framework of elastoplasticity,
the "global" response of the pile-soil system is treated in a manner similar to that for
the "local" response of an infinitesimally small soil element. The stresses and strains
for the soil element are substituted by the generalized forces (in 3-dimensional M-Q-
N space) and the corresponding displacements (θ–u–v) respectively. As with the
theory of elasto-plasticity, there are 5 main components in a macroelement
"assembly", associated with: (a) the foundation response at very small deformations

20
(elastic response). The elastic stiffness matrix has thus to be determined. (b) The
response of the foundation at very large deformations and at failure conditions. The
determination of the failure envelope in the generalized 3- dimensional M-Q-N space
is thus required. (c) The plastic flow rule that relates the incremental plastic
displacements of the foundation to its loading state at near failure conditions, (d) the
hardening rule that defines the transition from the elastic to the ultimate limit state,
and (e) the unloading-reloading rule in the case of cyclic loading.

2.3 Simplified Constitutive Model for RC Pile Section Behaviour

A Mohr-Coulomb based uniaxial stress-strain constitutive law is developed for

modeling the macroscopic behaviour of a RC circular pile section subjected to a
combined bending moment and axial force loading. Considering force equilibrium at
failure in the axial direction, one obtains:

( ) ( ) (1)

in which and are the compressive and tensile strength of the composite
(reinforced concrete) section, respectively:

(2)

where c, φ are the strength parameters of the Mohr-Coulomb model, namely the
cohesion and the internal friction angle. In Eq (1), and are the pile section
areas under compression and tension, respectively, defined as:

∫ √ ( ) (3)

(3)

21
 ∫ √ ( )

in which is the pile diameter, and is the abscissa (in a Cartesian coordinate
system) that defines the boundary between the zones of the section under
compression and tension, respectively. By applying moment limit equilibrium with
respect to the center of the pile section, the following equation is derived:

∫ √ ( )( )

(4)
∫ √ ( )( ) ( )

Eqs (1) and (4) form a nonlinear algebraic system. For a given pile diameter and a
known combination of bending moment–axial force at structural failure conditions,
there are three unknown variables: c, φ and . The aforementioned system is
solved with the use of a genetic algorithm–based optimization procedure,
implemented in MATLAB. The performed
optimization targets to a best fit on a predefined M–N failure envelope by
minimizing the relative root mean squared error (rRMSE) of the bending moment at
failure (the fitness function):

( ) √ ∑( ) (5)

in which is the bending moment computed by Eq (4), is the target bending

moment, and , the number of – pairs that define the failure envelope. It is
interesting to observe that the proposed simplified Mohr-Coulomb−based
constitutive model can be easily reduced to a Tresca with tension cutoff− based one,
by equating the compressive strength in Eq (2) with and setting the tensile
strength equal to the tension cut-off.
Fig 2.1 shows the M-N failure envelopes for a pile cross-section with a longitudinal
reinforcement ratio of As = 1.5% and for three different diameters (D = 0.8 m, 1 m
and 1.5 m). Comparison is given between the predictions of: (a) the proposed
optimization procedure, (b) the 3D FE analysis with PLAXIS, and (c) the fiber analysis
with the computer code USCRC (Esmaeily 2001). Fig 2.2 shows a 3D visualization of
the FE pile model. Observe the formation of a plastic hinge at the bottom of the pile
in terms of the incremental plastic shear strain.

22
2.4 Failure Envelope and Plastic Flow Rule for Piles

Problem Definition
The problem under consideration is that of a pile or a group of piles embedded in a
homogeneous cohesive soil of undrained shear strength Su. The pile / pilegroup is
subjected to a combined load of overturning moment, horizontal force and axial
force at the head / cap until complete failure.

Single Piles

2.4.1 Limit Equilibrium Analysis

Invoking Brom’s limit equilibrium theory (1964) for the ultimate lateral capacity of a
horizontally loaded free head pile (Fig 2.3), and assuming that the ultimate lateral
soil reaction per unit depth is approximated by the expression suggested by
Randolph and Houlsby (1984) and Broms (1964):

( ) (6)
{

the following analytical expression for the failure envelope is derived:

( ) (7)
√

In Eq (6), is the vertical effective stress and the effective specific unit weight of
the soil. Matlock (1970) stated that the value of was determined experimentally to
be 0.5 for a soft clay and about 0.25 for a medium clay. In Eq (7), is the bending
moment capacity of the pile which is a function of the axial load. The bending
moment capacity is fully mobilized at a certain depth through the formation of a
“plastic” hinge. , is a constant accounting for the distribution of the ultimate lateral
soil reaction along the pile. By taking into consideration all possible N-Q-M
combinations at the pile head and setting and √ , Eq
(7) is rewritten in the following general form:

23
 | ( )( ) ( )| || | |
(8)
| | | |

| |
{

in which is the axial bearing capacity of the pile subjected to tension or

compression:

( )( ( ) ) (9)

where is the capacity in tension (without the contribution of the pile tip) and
the compressive capacity (with due consideration to the pile tip).

2.4.2 Plastic Flow Rule

According to Brom’s theory the post-failure response of the pile is characterized by

the formation of a plastic hinge at a certain depth that acts as a rotation pole for the
above hinge rigidly deformed portion. With reference to Fig 2.3, the displacement to
rotation ratio can be approximated by:

(10)

in which , and , are the absolute and incremental plastic horizontal

displacements and rotations, respectively. Finally, is the depth to the plastic
hinge.

By differentiating Eq (7) with respect to the shear force Q and bending moment M,
respectively, and assuming an associative plastic flow rule, one obtains after some
algebra:

(11)

24
in which is the value of at failure conditions. Thus, in the framework of a limit
equilibrium analysis of the lateral capacity of a single pile, the plastic flow rule is by
definition of the associative type!

The undoubtedly very interesting findings about the plastic flow rule and failure
criterion are about to be verified in the following section by 3D finite element
analysis.

2.4.3 Finite Element Analysis with PLAXIS

A 16 m long pile with a diameter of 1 m embedded in a homogeneous cohesive soil,

is analysed. Both the pile and the soil are modeled with 10-node tetrahedral
elements. The size of the finite element model is 1.3L x 1.3L in plan view with a
depth of 1.5 L (where L is the length of the pile), carefully weighting the effect of the
boundaries on the response of the pile and the computational time. Zero-
displacement boundary conditions prevent the out of plane deformation at the
vertical sides of the model, while the base is fixed in all three directions. Special
interface elements are placed between the pile and the soil, thus allowing slippage
and gapping to occur. For the total stress analysis under undrained conditions, soil
behaviour is described by the Mohr- Coulomb model with c = Su = 50 kPa, φ = ψ (the
dilation angle) = 0°, specific weight of γ = 18 kΝ/m3, elasticity modulus of Es = 25000
kPa and Poisson’s ratio of v = 0.45. Based on the macroscopic constitutive law for
reinforced concrete circular pile sections (in section 3), the behaviour of the pile was
modeled via a Mohr-Coulomb failure criterion with c = 15262 kPa, φ = 0ο, tension
cut-off σt = 7534 kPa and Elasticity modulus E = 30 GPa, corresponding to a
longitudinal reinforcement ratio of As = 1.5 %. Fig 2.4 compares the failure envelopes
in M-Q-N space as predicted by the analytical expression of Eq (8) and calculated by
the FE models. The hypothesis of plastic flow rule implied by the limit equilibrium
analysis is verified in Fig 2.5 through comparison with finite element analysis results.
The discontinuity in plastic hinge position observed at a load angle of |ω| ≈ 50°,
where the failure envelope reduces to a straight line of constant bending moment, is
consistent with associative plastic flow rule. Indeed, for |ω| ≥ 50°, perpendicularity
of the incremental plastic pile deflection at this point implies failure under pure
rotation and zero lateral displacement. The only possible failure mechanism that
satisfies this geometrical constraint is the formation of a plastic hinge at the head of
the pile.

25
2.5 Macroelement Modeling

The model is formulated in the framework of classical elastoplasticity, and combines

features of: (a) the bounding surface plasticity, (b) the critical state concept, and (c) a
hardening evolution law and unloading-reloading rule of the modified Bouc-Wen
type. According to this formulation the tangent elastoplastic stiffness matrix that
relates the incremental force vector to the incremental displacement vector, is given
by:

[ ( ) ] (12)

in which is the elastic stiffness matrix of the pile, and account for the
failure surface and plastic flow rule, respectively, and and describe the
hardening law and unloading-reloading rule. The terms in matrices and are
functions of the dimensionless hardening parameter ζ, which is of the Bouc-Wen
type (Gerolymos and Gazetas, 2005). Figs 2.6 and 2.7 presents numerical examples
of the macroelement model for a pile subjected to combined loading of axial force,
horizontal force and overturning moment at its head. Comparison is given with
results from finite element analysis with code PLAXIS. The properties of the pile and
its supporting soil are provided in section 2.4.3.

26
Figures

27
Figure 2.1. Comparison of M-N failure envelopes for RC pile sections, computed from
the Fiber analysis and from PLAXIS and predicted by the proposed optimization
procedure for three pile diameters: (a) D = 0.8 m, (b) D = 1 m, (c) D = 1.5 m. (d)
Comparison of the bending moment-curvature curve calculated from the fiber
analysis and the FE models (PLAXIS) for D = 1 m.

28
Figure 2.2. (a) The FE model (PLAXIS) for Figure 2.3. Failure mode of a laterally
RC pile section analysis, (b) the deformed loaded free-head pile embedded in
FE mesh at failure, (c) contours of the cohesive soil according to Brom’s theory.
incremental plastic shear strains at
failure denoting the formation of a
plastic hinge at the bottom of the pile.

Figure 2.4. Comparison of the failure envelopes for a pile in cohesive soil, calculated
by PLAXIS and predicted by the analytical expression [Eq (8)] for 5 different factors of
safety to vertical loading Fsv. The pile has a diameter of D = 1 m and a longitudinal
reinforcement ratio of As = 1.5 %. Tension (extraction of the pile) is denoted with
negative values.

29
Figure 2.5. Finite element verification of the associative plastic flow rule for a wide
range of load combinations. Observe that the depth to the plastic hinge hp increases
for increasing negative load angles *M / Q = ω, ω* = (Qy / My) ω+ reaching a
maximum value at the vertex of the failure envelope (at approximately ω = -50°). For
load angles greater than |ω| > 50° , the plastic hinge moves violently from its
deepest location to the head of the pile.

30
Figure 2.6. Comparison of horizontal force Figure 2.7. Comparison of overturning
versus horizontal displacement curves at the moment versus rotation curves at the head
head of the pile, predicted by the of the pile, predicted by the macroelement
macroelement and calculated from the finite and calculated from the finite element
element analysis (PLAXIS) for the following analysis (PLAXIS) for the following load
load combinations: (a) Pure shear loading and combinations: (a) Pure moment loading
Fsv = inf, (b) combined shear force-overturning and Fsv = 5, (b) combined shear force-
moment for a load angle of ω = - 70° and Fsv = overturning moment for a load angle of ω=
inf -36° and Fsv = -2 (tension)

31
 Chapter 3
Internal Force Diagrams Computation

33
3 Internal Force Diagrams Computation

3.1 PLAXIS approach

Although PLAXIS can calculate internal force diagrams of elastic piles, this is not
the case with non-linear piles. In elasticity, the internal forces can be calculated
by putting a slender beam along the shaft of the pile with a Young modulus (E)
approximately 10.000 times smaller of the stiffness of the pile, so as not to
affect the pile response in loading. By completing the calculations, the beam
returns the force components in relation with the strain / curvature
components. This relation is defined according to Bernoulli beam theory:

where A: Beam cross section area

E: Young’s Modulus in axial direction
I2: Moment of inertia against bending around the second axis
I3: Moment of inertia against bending around the third axis
and k is the shear correction factor. The shear modulus G is taken as G .

The above-mentioned force components are multiplied by the factor m

so as to get the real diagrams of axial force, shear force and bending moment.
However, the linearity does not apply for loads near the failure of pile, because
in this case emerges severe plasticity which leads to a reduction of the elastic
modulus and the above-mentioned equations cannot explain the actual pile
response. As a result, emerges the need of a tool to compute these forces,
analyze them and propose ways of better pile design.
In this thesis a model of numerical integration through MATLAB programming
language is proposed. The procedure that is followed takes as input data the
stresses which are derived from the experiment simulation results and through a
series of commands, the structural force diagrams are extracted.
The concept is to use the stresses of the Gauss points and by interpolating them
in a circular area of same depth, define an area of influence of each interpolated
point, in addition integrate them numerically and extract the results at each p ile
depth.

3.2 Voronoi decomposition

The region of influence of each interpolated Gauss point is defined by the

Voronoi tessellation method. The finite set of points are given in the cyclic plane
and each site p k is simply a unique point and its corresponding Voronoi cell

34
Pk consisting of every point whose distance to p k is less than or equal to its
distance to any other site. Each such cell is obtained from the intersection of
half-spaces, and hence it is a convex polygon. The segments of the Voronoi
diagram are all the points in the plane that are equidistant to the two nearest
sites. The Voronoi vertices (nodes) are the points equidistant to three (or more)
sites. However, a problem arises due to the fact that the region of influence of
the points across the circumference extends to infinity. The proposed solution to
this problem is to regard pseudo-points adjacent to the points of the
circumference, along the radius of circle and on the outer side. These points are
neglected afterwards in the calculations.

X=CO(:,1);
Y=CO(:,2);
[v,c]=voronoin([X(:) Y(:)]);

3.3 Numerical Integration

In addition to Voronoi decomposition, the area of the unique regions is acquired

so as to proceed to the numerical integration. For the case where only
symmetric loading x-wise is applied to the head of the pile, internal forces are
calculated according to the expressions:

∑ ( ) ( )
(1)

∑ ( ) ( )( ( )) (2)

∑ ( ) ( ) (3)

Where: N is the axial force, M is the bending moment, Q is the shear force and n
is the number of total Voronoi regions. and are the stresses of the
elements of the pile, the sign of which is conventionally defined as in figure 3.3.
Xc is the x-component of the center of a pile intersection. The reference point
according to which moments are determined is the Xc, so as to avoid measuring
any additional moment that may be caused due to the eccentricity of the normal
stresses distribution. Two approaches were made to estimate the force values of
the pile. One in 2-Dimensional area and one in 3-Dimensional space.

35
3.4 Slice Method

This approach attempted to take advantage of the automatic interpolation

PLAXIS made in every required depth section. The interpolation to the specific
depth is obtained in the finite element formulation from the interpolation of the
nodal normals with the shape functions. The command script reads a five-
column matrix, every row of which is representing the X position, Y position, Z
position, the σzz stress and σzx stress. This can be easily modified in order to take
into account moment and shear forces in the perpendicular direction. However,
this procedure produces different stress values to same points due to the fact
that two or more adjacent Gaussian points are interpolated to the same
intersection point. These interpolated values are averaged in order to have
unique points and their values, and then sorted to abbreviate the calculation
time.

P=xlsread('Plaxis_Mat.xlsx','Plaxis_Mat','A3:E362');
P(:,3)=[];
CO_plaxis=unique(P(:,1:2),'rows','sorted');
last_plaxis = length(CO_plaxis);
X_plaxis=CO_plaxis(:,1);
Y_plaxis=CO_plaxis(:,2);
sigma_plaxis=zeros(size(X_plaxis));
shear_plaxis=zeros(size(X_plaxis));
for i=1:last_plaxis
k=0;
sum1=0;
sum2=0;
for j=1:length(P)
if and(P(j,1)==CO_plaxis(i,1),P(j,2)==CO_plaxis(i,2))
k=k+1;
sum1=sum1+P(j,3);
sum2=sum2+P(j,4);
end
end
sigma_plaxis(i,1)=sum1/k;
shear_plaxis(i,1)=sum2/k;
end

3.5 Interpolation

The automatically produced section-points of PLAXIS, may not be well-handled,

because, for instance, of the bad combination of the circumferential points with
the respective outer Voronoi pseudo-points, or because of probable inadequate
distribution of stresses throughout the pile intersection. What is proposed is to
perform interpolation on the 2-D scattered data set, in order to pass the values
of stresses from the sample values (sigma_plaxis) at the given point locations
(X_plaxis,Y_plaxis) to any query point (x,y) and produce a linearly interpolated

36
value of stress (sigma). To achieve the interpolation, it must be defined a grid
corresponding to the polar coordinates of the query points. This grid is
afterwards converted to a Cartesian-coordinate set of points, for better
handling. Regarding the points very close to the circumference or on it, which
are not enclosed to the area defined by the Gauss points and cannot be
interpolated, it is chosen an extrapolation method according to which, these
points acquire the stress value of their nearest Gauss point. Although this
procedure may sound approximative or inaccurate, extrapolation is applied only
to a minor set of points with only slightly different stress value of their adjacent
integration points, that no significant change of the structural forces is detected.

r=linspace(0,(Xmax-Xmin)/2,20);
theta=linspace(0,2*pi,40);
[r,theta]=meshgrid(r,theta);
x=(Xmax+Xmin)/2+r.*cos(theta);
y=(Ymax+Ymin)/2+r.*sin(theta);
F_sigma=scatteredInterpolant(X_plaxis,Y_plaxis,sigma_plaxis, ...
'linear','nearest');
sigma=F_sigma(x,y);
F_shear=scatteredInterpolant(X_plaxis,Y_plaxis,shear_plaxis, ...
'linear','nearest');
shear=F_shear(x,y);

However, the grid of the interpolated points is not well-handled; hence the
coordinates and their respective values are reshaped into the more convenient
form of vectors. In addition, the double coordinate values are deleted and the
remaining are sorted in increasing order, in respect to the X coordinate.
Subsequently, the external boundary coordinates are added to the Nх2
coordinate matrix and then, the area of the polygon specified by the vertices in
the vectors X and Y of each Voronoi polygon is created. The forces can be
calculated according to the (1) - (3) equations.

37
x=reshape(x,[],1);
y=reshape(y,[],1);
sigma=reshape(sigma,[],1);
shear=reshape(shear,[],1);
Ptemp=[x,y];
[COtemp,ia,ic]=unique(Ptemp,'rows','sorted');
sigmatemp=zeros(size(ia));
for i=1:length(ia)
sigmatemp(i)=sigma(ia(i));
end
sigma=[zeros(extsize,1); sigmatemp];
sheartemp=zeros(size(ia));
for i=1:length(ia)
sheartemp(i)=shear(ia(i));
end
shear=[zeros(extsize,1); sheartemp];
CO=[bounds;COtemp];
last = length(CO);
X=CO(:,1);
Y=CO(:,2);
[v,c]=voronoin([X(:) Y(:)]);
figure;
voronoi(X,Y)
plabels = arrayfun(@(n) {sprintf('P%d', n)}, (1:last)');
Hpl = text(X, Y, plabels, 'FontWeight', ...
'bold', 'HorizontalAlignment','center', ...
'BackgroundColor', 'none');
dA=zeros(last-extsize,1);
for i=(extsize+1):last
dA(i)=polyarea(v(c{i},1),v(c{i},2));
end
A=sum(dA);
title(['Area = ' num2str(A)]);
dN=zeros(last-extsize,1);
for i=(extsize+1):last
dN(i)=dA(i)*sigma(i);
end
N=sum(dN);
coeff=polyfit(X((extsize+1):last),sigma((extsize+1):last),1);
x0=coeff(1,2)/(-coeff(1,1));
dM=zeros(last-extsize,1);
for i=(extsize+1):last
dM(i)=dA(i)*sigma(i)*((Xmax+Xmin)/2-X(i));
end
M=sum(dM);

dQ=zeros(size(X_plaxis));
dQ((extsize+1):last)=dA((extsize+1):last).*shear((extsize+1):last);
Q=sum(dQ);
NMQ=[N;M;Q]

38
3.6 Volume method

There are two main disadvantages of using separate intersections in two

dimensional space. First, the multiple interpolations that have to be done both
manually by the commands and automatically with PLAXIS, return results with
low accuracy and considerable divergence. In addition, the internal force values
have to be calculated in each depth separately and then the diagrams be
formed.
Regarding this issue, there is proposed a different approach which takes into
account the stresses of the whole pile, directly from the Gauss points. These
stresses are interpolated at each depth that is queried and subsequently, the
former procedure of the separate intersection approach is followed. In this case,
the Voronoi decomposition is also made after the interpolation in the plain.
The Coordinate matrix is now Nx3 and the third column is related with the depth
. The procedure is similar to the one in 2-D approach, but in this case the grid is
a three dimensional matrix and the interpolation is done in space.
r=linspace(0,(Xmax-Xmin)/2,25);
theta=linspace(0,2*pi,80);
zeta=linspace(Zmax,Zmin,abs(Zmin-Zmax)+1)';
zeta1=zeta;
[r,theta,zeta]=meshgrid(r,theta,zeta);
xgrid=(Xmax+Xmin)/2+r.*cos(theta);
ygrid=(Ymax+Ymin)/2+r.*sin(theta);

F_sigma = scatteredInterpolant(CO_plaxis,sigma_plaxis, ...

'linear','nearest');
sigmagrid = F_sigma(xgrid,ygrid,zeta);
F_shear = scatteredInterpolant(CO_plaxis,shear_plaxis, ...
'linear','nearest');
sheargrid = F_shear(xgrid,ygrid,zeta);

Regarding the internal forces generation, it is done at each predefined depth of

pile, inside a for-loop. In each loop, the forces are calculated according to the
procedure defined in 2-D approach.

39
NMQ=zeros(3,size(zeta,3));
for I=1:size(zeta,3)
xloop=reshape(xgrid(:,:,I),[],1);
yloop=reshape(ygrid(:,:,I),[],1);
sigmaloop=reshape(sigmagrid(:,:,I),[],1);
shearloop=reshape(sheargrid(:,:,I),[],1);
Ploop=[xloop,yloop];
[COloop,ia,ic]=unique(Ploop,'rows','sorted');
sigmalooptemp=zeros(size(ia));
shearlooptemp=zeros(size(ia));
for i=1:length(ia)
sigmalooptemp(i)=sigmaloop(ia(i));
shearlooptemp(i)=shearloop(ia(i));
end
sigma=[zeros(extsize,1); sigmalooptemp];
shear=[zeros(extsize,1); shearlooptemp];
CO=[bounds;COloop];
last = length(CO);
X=CO(:,1);
Y=CO(:,2);
[v,c]=voronoin([X(:) Y(:)]);
plabels = arrayfun(@(n) {sprintf('P%d', n)}, (1:last)');
Hpl = text(X, Y, plabels, 'FontWeight', ...
'bold', 'HorizontalAlignment','center', ...
'BackgroundColor', 'none');
dA=zeros(last-extsize,1);
for i=(extsize+1):last
dA(i)=polyarea(v(c{i},1),v(c{i},2));
end
A=sum(dA);
title(['Area = ' num2str(A)]);
dN=zeros(last-extsize,1);
for i=(extsize+1):last
dN(i)=dA(i)*sigma(i);
end
N=sum(dN);
coeff=polyfit(X(extsize+1:last),sigma(extsize+1:last),1);
x0=coeff(1,2)/(-coeff(1,1));
dM=zeros(last-extsize,1);
for i=(extsize+1):last
dM(i)=dA(i)*sigma(i)*((Xmax+Xmin)/2-X(i));
end
M=sum(dM);
dQ=zeros(last-extsize,1);
dQ((extsize+1):last)=dA((extsize+1):last).*shear ...
((extsize+1): last);
Q=sum(dQ);
NMQ(:,I)=[N;M;Q];
end

40
3.7 Validation

The validity of the results of the above-mentioned calculations is examined by

comparing the internal force diagrams of beam element in elasticity with linear
behavior of soil and pile and the respective diagrams that are extracted from the
matlab code. The results are also validated by semi-analytical expressions that
give the moment of a pile on a given depth according to the distribution of the
normal stresses. For the simulation, is used a monopile which is subjected
exclusively to lateral loading in x direction.

Semi - analytical expression

We assume a circular pile intersection in depth with polar coordinates (r,θ), as

in figure 3.2.

𝛭 ( ) ∫
and d d

( ) ∗ ∫∫

Because of the elastic behavior and the absence of axial loading, it is fair to
assume linear stress distribution with zero value at the pile center.
( ) ( )
𝑅 𝑅
Thus,

( )
( ) ∗ ∫∫
𝑅

( )
( ) ∗ ∫∫
𝑅

⇒ ( ) 𝑅 ( )

Where
The results of the validation are shown in figures 3.25 – 3.27.

41
3.8 Result visualization

N-M-Q diagrams

The aforementioned results are eventually visualized to get also a qualitative

indication of their validity and a general perspective of the pile response under
static loading. Thus, it is possible a practical comparison of the structural forces
that are extracted from the proposed method with the respective ones from
alternative methods. These diagrams are handled for constructional purposes
regarding the design of the pile and the pile enforcement. The presentation of
the diagrams is chosen to form an area of three diagrams in a row.

%Visualizations
subplot(1,3,1);
plot(NMQ(1,:),zeta1,'-b*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Axial Force (N)', 'FontSize', 20);
xlabel('(kN)');
ylabel('Depth (m)');
set(gca,'XAxisLocation','top');
set(gca,'YAxisLocation','right');
hold off
subplot(1,3,2);
plot(NMQ(2,:),zeta1,'-r*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Bending Moment (M)', 'FontSize', 20);
xlabel('(kNm)');
ylabel('Depth (m)');
set(gca,'XAxisLocation','top');
set(gca,'YAxisLocation','right');
hold off
subplot(1,3,3);
plot(NMQ(3,:),zeta1,'-g*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Shear Force (Q)', 'FontSize', 20);
xlabel('(kN)');
ylabel('Depth (m)');
set(gca,'XAxisLocation','top');
set(gca,'YAxisLocation','right');
set(gcf,'Color',[0.4,0.4,0.4],'Toolbar','none')
set(gcf, 'Position', get(0,'Screensize'));
hold off

42
3.9 Stress distribution

In conjunction with the internal force diagrams, the normal and shear stress
distribution in each depth are visualized. Matlab has not the option to plot stress
distribution in a cyclic area. Thus, the proposed solution is to create a squared
intersection with edge length equal to the pile diameter and grid density
unvarying to the Voronoi grid density. The stress value of each grid on the
square is acquired by bilinear, 2-D interpolation of the p k sites of the Voronoi
grid. For presenting reasons, the intersections shown on the 3-D figures 3.22 and
3.23 are at two meter distance.

[x_rec, y_rec, z_rec] =

meshgrid(linspace(Xmin,Xmax,100),linspace(Ymin,Ymax,100)…
,linspace(Zmax,Zmin,abs(Zmin-Zmax)/2+1));
sigma_rec = griddata(xgrid,ygrid,zeta,sigmagrid,x_rec,y_rec,z_rec);
shear_rec = griddata(xgrid,ygrid,zeta,sheargrid,x_rec,y_rec,z_rec);

figure;
slice(x_rec,y_rec,z_rec,sigma_rec,[],[],[Zmax:-2:Zmin]);
axis ([10 11 10 11 -6 0]);
title('Normal Stress (kPa)');
set(gcf,'Color',[0.4,0.4,0.4])
colorbar
zoom (3)
pan on
set(gcf, 'Position', get(0,'Screensize'));
figure;
slice(x_rec,y_rec,z_rec,shear_rec,[],[],[Zmax:-2:Zmin]);
axis equal;
title('Shear Stress (kPa)');
set(gcf,'Color',[0.4,0.4,0.4])
colorbar
zoom (18)
pan on
set(gcf, 'Position', get(0,'Screensize'));

43
44
Figures

45
 Area = 0.79759
11.4

11.2

11 P8
P84 P7
P143 P6
P185
P9
P56 P88 P142 P181 P5
P213
P10
P40
P63 P95 P141 P174 P207 P229 P4
P69 P99 P140 P170 P200
10.8 P47 P222
P52 P74 P105P139P164 P196 P217
P11
P31 P60 P82 P107 P187 P209 P238
P3
P36 P138P162 P233
P42 P67 P91 P179 P202 P227
P49 P137
P116 P153 P220
P12
P28P33 P58 P78 P101 P168P191 P211 P2
P241
10.6 P38P45 P71 P86P110
P120P149 P183P198
P136 P236
P231
P103 P160
P54P65 P80P97 P166P189P204 P224
P124
P135
P145
P122
P76P93P114
P118 P147
P151P172P193 P215
P112 P176
P155
P157
P13
P26P29P34P43P50P62P73P90P109
P111P134
P159
P156P178
P195
P206P218
P225
P234P239
P242
P25
P1
P64P75P92P113
P96P117
P121 P154
P150
P146
P123
P133
P144
P175
P171P192
P203
P44P53 P79 P102 P165P188 P214
P223
10.4 P27P32
P14
P37 P70 P85P108
P119 P158P182P197
P148
P132 P230
P235
P240
P24
P57 P77 P100P115 P167P190 P210
P152
P48 P131 P219
P41 P66 P89 P177 P201 P226
P35 P59 P81 P106P130P161 P232
P30
P15 P186 P208 P237
P23
P51 P72 P104P129P163 P194 P216
10.2 P46
P68 P98 P128 P169 P199
P221
P39
P16 P94 P127 P173 P205 P228 P22
P61
P55
P17 P87 P126 P180 P212
P21
P83
P18 P184
P20
10 P125
P19

9.8
9.8 10 10.2 10.4 10.6 10.8 11 11.2 11.4

Figure 3.1. Intersection of a pile at specific depth. Pk are the Voronoi regions to which
the pile is intersected.

Figure 3.2. Polar coordinates of pile Figure 3.3. General three dimensional
intersection stress points coordinate system and sign convention for
stresses

46
Figure 3.4. Instance of lateral loading of the monopile. This type of loading is examined
below

Figure 3.5. Pile intersection along the depth in Figure 3.6. PLAXIS table, from
Slice Method. The element nodes must which, the normal and shear
coincide with the pile depth at the specific stresses at each specific point are
intersection extracted

47
 11 11

10.5
10.5
r

y
10
10
x 10.5 11
10
10 10.5 11

Figure 3.7. Node distribution at the pile Figure 3.8. Node distribution at the pile
intersection of the non-elaborated intersection after the implementation of
data. Many nodes have varied values specific grid pattern to the stress points
of stresses due to the interpolation of
different Gauss points

4 4
x 10 x 10
1.5 1.5

1 1

0.5 0.5
normal stress (kPa)
normal stress (kPa)

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5
10 10.2 10.4 10.6 10.8 11 10 10.2 10.4 10.6 10.8 11
X (m) X (m)

Figure 3.9. Normal stress distribution, in Figure 3.10. Normal stress distribution, in
elasticity, of the non-elaborated data elasticity, of the interpolated data

48
 1800 1800

1600 1600

1400 1400
shear stress (kPa)

shear stress (kPa)

1200 1200

1000 1000

800 800

600 600

400 400

200 200
10 10.2 10.4 10.6 10.8 11 10 10.2 10.4 10.6 10.8 11
X (m) X (m)

Figure 3.11. Shear stress distribution, in Figure 3.12. Shear stress distribution, in
elasticity, of the non-elaborated data elasticity, of the interpolated data

4
x 10 x 10
4
3 2

2
1

1
0
normal stress (kPa)

normal stress (kPa)

0
-1
-1
-2
-2

-3
-3

-4
-4

-5 -5
10 10.2 10.4 10.6 10.8 11 10 10.2 10.4 10.6 10.8 11
X (m) X (m)

Figure 3.13. Non-linear stress distribution Figure 3.14. Non-linear stress

of normal stresses, of the non-elaborated distribution of normal stresses, of the
data interpolated data

49
 6000 6000

5000 5000

4000 4000

shear stress (kPa)

shear stress (kPa)

3000 3000

2000 2000

1000 1000

0 0

-1000 -1000

-2000 -2000
10 10.2 10.4 10.6 10.8 11 10 10.2 10.4 10.6 10.8 11
X (m) X (m)

Figure 3.15. Non-linear stress distribution Figure 3.16. Non-linear stress

of shear stresses, of the non-elaborated distribution of normal stresses, of the
data interpolated data

Normal Stress (kPa) x 10

4 Shear stress (kPa)
1 1600

10.9 0.8 10.9

1400
10.8 0.6 10.8

10.7 0.4 10.7 1200

10.6 0.2 10.6

Y-axes

Y-axes

1000
10.5 0 10.5

-0.2
10.4 10.4 800
-0.4
10.3 10.3
-0.6 600
10.2 10.2
-0.8
10.1 10.1 400
-1
10.2 10.4 10.6 10.8 11 10.2 10.4 10.6 10.8 11
X-axes X-axes

Figure 3.17. Contour of normal stress Figure 3.18. Contour of shear stress
distribution of the interpolated data, in distribution of the interpolated data, in
elasticity elasticity

50
 Normal Stress (kPa) x 10
4 Shear stress (kPa)
11

1 10.9 5000

10.8 10.8
0 4000
10.7

10.6 10.6 3000

-1
Y-axes

Y-axes
10.5
2000
-2
10.4 10.4
1000
-3 10.3

10.2 10.2
0
-4
10.1
-1000
10
10.2 10.4 10.6 10.8 10.2 10.4 10.6 10.8 11
X-axes X-axes

Figure 3.19. Contour of non-linear normal Figure 3.20. Contour of non-linear shear stress
stress distribution of the interpolated data distribution of the interpolated data

0 0

-1 -1

-2 -2

Figure 3.21. Distribution of the non- Figure 3.22. Distribution of the stress points
elaborated stress points (Gauss points) in (Gauss points) in the pile volume after the
the pile volume (only the first 2 meters of spatial interpolation to specific depths (only
the pile are shown) the first 2 meters of the pile are shown)

51
Figure 3.23. Normal stress distribution at Figure 3.24. Shear stress distribution at each
each pile depth(only the first 2 meters of pile depth(only the first 2 meters of the pile are
the pile are shown) shown)

Figure 3.25. Total Gauss points along the Figure 3.26. Stress points after elaboration.
pile as extracted from PLAXIS Internal Forces are computed at each depth
according to the interpolated data

52
 Q (kN) M (kNm)
-500 0 500 1000 1500 2000 -500 0 500 1000 1500 2000
0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10

-12 -12 Plaxis

-14 Plaxis -14 Method of slices

-16 Method of slices -16 Volume method

-18 -18
Volume method Analytical
Expression
-20 -20

Figure 3.27. Shear Force. Comparison of the Figure 3.28. Bending moment. Comparison
two methods with the results derived from of the two methods with the results derived
PLAXIS, in elasticity from PLAXIS, in elasticity and the semi-
analytical expression

N (kN)
-400 -300 -200 -100 0 100 200 300
0

-2
Plaxis

-4 Method of slices

Volume Method
-6

-8

-10

-12

-14

-16

-18

-20

Figure 3.29. Axial Force. Comparison of the two methods with the PLAXIS results, in
elasticity. The difference between the proposed methods and the PLAXIS results may
be attributed to the assumption PLAXIS makes which regards the pile as beam. Thus,
in order to define border conditions for the pile, considers the pile footing as free
edge with zero axial force. In reality, the pile footing commits axial force due to the
difference between the specific weight of the reinforced concrete of the pile and that
of the soil.

53
 Axial Force (N) Bending Moment (M) Shear Force (Q)
(kN) (kNm) (kN)
-300 -200 -100 0 100 -500 0 500 1000 1500 2000 -500 0 500 1000 1500
0 0 0

-2 -2 -2

-4 -4 -4

-6 -6 -6

-8 -8 -8
Depth (m)

Depth (m)

Depth (m)
-10 -10 -10

-12 -12 -12

-14 -14 -14

-16 -16 -16

-18 -18 -18

-20 -20 -20

Figure 3.30. Internal Force diagrams (N,M,Q) as extracted at each pile depth
according to the Volume Method

54
 Chapter 4

Application to Single Pile

55
4 Application to Single Pile

4.1 Limit Equilibrium Approach

The case of a free‐head flexible pile embedded in clay with constant undrained
shear strength Su is studied. In this case the soil resistance is
, 4.1

where, as mentioned before:

(Matlock 1970) 4.2

And A λ , otherwise

From static equilibrium:

4.3

4.4

From 3.3:

4.5

And by applying (4.5) in (4.4):

( ) ⇒

( ) 4.6
√

56
By setting:

√
equation (3.6) becomes

( ) 4.7

By taking into consideration all possible N-Q‐M combinations at the pile head
the failure envelope for a flexible pile embedded in clay with constant undrained
shear strength (4.8-4.10):

| ( )( ) ( )| || | |

| | | |

| |
{

where,

( )( ( ) ) 4.11

Nc , Nt the ultimate compressive and tensile capacity respectively.

Assuming an associated flow rule (in which the plastic potential function g
coincides with the yield function f) the plastic displacement u pl and the plastic
rotation φpl give:

4.12

57
confirming that the incremental plastic displacement vectors at the point of failure
are normal to the yield locus.

4.2 Finite Element Verification

The proposed failure envelopes and the new method for the calculation of the
structural forces are checked against three‐dimensional numerical analysis for
flexible pile and pile-groups embedded by using the finite element code Plaxis 3D.

4.2.1 Static pushover tests

Considering that the foundation supports a 1-DoF oscillator, one expects that radial
loading paths on the M-Q plane are applied in the system. Through a series of force-
controlled analyses the failure envelope is ultimately determined. Prior to that, the
foundation has undergone vertical loading N to a fraction χ = N/Nu of its ultimate
capacity. [Cremer, Pecker, Davenne 2001; Gouvernec 2004; Gajan, Kutter, Phalen,
Hutchinson, Martin 2005].

4.2.2 Static pushover tests

The steps followed in our numerical experiments represent the actual conditions in
the field. The soil undergoes geostatic loading and then a part of the soil is replaced
by the foundation, on which a vertical load N is applied increasingly till a specified
value of χ = Ν/Νu is reached. Afterwards, the vertical load is kept constant and a
combination of horizontal force and moment is applied at the head of the pile till the
complete failure of the system. Apparently, this implies the state in which no further
lateral loading can be undertaken. The above procedure is repeated for various
factors of safety against vertical loading and for various radial loading paths. Our aim
is to extract the ultimate capacities under pure moment Mu and pure horizontal
force Qu, and then sweep the M-Q plane so that a cross-section of the failure
envelope is revealed. Repeating this procedure from the Ultimate Axial Compression
Capacity to the ultimate Axial Tension Capacity the total 3D Failure Envelope in M-
QN space is designed.

58
4.3 Finite Element Modelling

A 16 m long pile with 1 m diameter is embedded in the soil. The distance from the
pile tip to the bottom of the model is 6 m. Figure (4.5) depicts the finite element
discretization of the problem. Approximately 48000 elements were used for each
analysis. The soil is modeled with 10‐node tetrahedral elements while the pile is
modeled as a soil volume calibrated with the previously macroscopic hardening Soil
model approach to simulate the behavior of a circular concrete pile with As=1.5% . A
sensitivity analysis for the lateral boundaries is carried out to ensure the accuracy of
the model, placing them finally at the distance of 0.6L. The selected Soil is Clay with
constant with the depth Undrained Shear Strength Su=50 kPa, specific weight γ=20
kN/m3 and Es=25000 kN/m and its behavior is described by the Mohr-Coulomb
Model. The poisson’s ratio is v=0.45 while the angle of friction is φ=0° to simulate
undrained water conditions. The pile has an elasticity Modulus of Ec=30*10^6 KN/m,
a poisson’s ratio v=0.2 and a specific weight practically zero (γ=0.01 KN/m2) to
ensure that the derived ultimate loads are the total ones, while cohesion is chosen
to be c=15262 kPa, the angle of friction φ=0° and tension cut-off strength equal to
7534 kPa in order to capture the correct pile behavior. An Interface is used between
the pile and the soil enabling gapping and slippage with a friction coefficient R=1.

4.4 Results

Figure 4.1 displays the failure envelope of concrete pile, as defined from the
equation 4.7. Figure 4.2 shows the combinations of moment and shear force at
failure, which were chosen for the study of the pile forces. The influence of the
vertical load can clearly be captivated in Fig. 4.3 where the maximum capacities
magnify by the increase of the axial load. Figures 4.8 – 4.25 show the structural
forces as extracted from MATLAB code, for combinations at failure and for
characteristic failure mechanisms. The effect of axial force to the moment capacity
and the moment distribution along the pile is examined afterwards, in figures 4.26 –
4.32. Firstly, the case of single pile embedded in soft clay is studied, subjected to
different axial loading, which results to different factors of safety.

59
60
Figures

61
 1.5

0.5

Q/Qy
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

-0.5

-1

-1.5
M/My
Figure 4.1. Failure envelope for specific axial force, for a circular concrete pile with
As=1.5%

1.5

a (0,Mu)
1

b (√2/2Qu,0.5Mu)
0.5
M/My

c (Qu,0) Q/Qy
0
0 0.5 1 1.5 2

-0.5
d (1.23Qu,-0.5Mu)

-1

e (0.75Qu,-Mu)
-1.5

Figure 4.2. From the y’y symmetric failure envelope, 5 combinations of Moment and
Shear force were chosen, that represent characteristic areas of the failure zones

62
 31

26

21
Axial (MNm)

16

11

1
0 1000 2000 3000
-4
Moment (kNm)
-9
Figure 4.3. Moment – Axial force interaction diagram, for cylindrical concrete pile
with As=1.5%. The dependence of the moment capacity with respect to the pile axial
force is represented

63
 A
B Plastic Hinge
1.5
C
h
1.0
A
0.5 B
M/My

0 D
C
-0.5 D
-1.0

-1.5 0 0.5 1.0 1.5

Q/Qy
Figure 4.4. Failure mechanisms of single piles with respect to the different Moment –
Shear force combinations

1m
H

22 m

Figure 4.5. The finite element model of single, flexible pile in clay

64
Figure 4.6. Vertical intersection of pile under horizontal loading. This figure depicts
clearly the three types of non-linearity and their interaction as a system: The soil non-
linearity, the pile plastification and the geometric non linearities

Figure 4.7. Pile under horizontal loading. The three figures show respectively the
deformed mesh of pile in failure, the pile deviatoric strains (Δγs) which show the
plastic hinge region, and the pile plastic points i.e. the plasticized Mohr – Coulomb
points (red) and the tension cut-off points (white)

65
Figure 4.8. Structural forces of single pile with SFv=2 at failure combination: (√ ⁄ )

Figure 4.9. Structural forces of single pile with SFv=2 at failure combination: ( )

66
Figure 4.10. Structural forces of single pile with SFv=2 at failure combination: ( )

Figure 4.11. Structural forces of single pile with SFv=2 at failure combination: ( )

67
Figure 4.12. Structural forces of single pile with SFv=1.25 at failure combination: (√ ⁄ )

Figure 4.13. Structural forces of single pile with SFv=1.25 at failure combination: ( )

68
Figure 4.14. Structural forces of single pile with SFv=1.25 at failure combination: ( )

Figure 4.15. Structural forces of single pile with SFv=1.25 at failure combination: ( )

69
Figure 4.16. Structural forces of single pile with SFv=∞ at failure combination: (√ ⁄ )

Figure 4.17. Structural forces of single pile with SFv=∞ at failure combination: ( )

70
Figure 4.18. Structural forces of single pile with SFv=∞ at failure combination: ( )

Figure 4.19. Structural forces of single pile with SFv=∞ at failure combination: ( )

71
Figure 4.20. Structural forces of single pile with SFv=∞ at failure combination: ( )

Figure 4.21. Structural forces of single pile with SFv= -2 at failure combination: (√ ⁄ )

72
Figure 4.22. Structural forces of single pile with SFv= -2 at failure combination: ( )

Figure 4.23. Structural forces of single pile with SFv= -2 at failure combination: ( )

73
Figure 4.24. Structural forces of single pile with SFv= -2 at failure combination: ( )

Figure 4.25. Structural forces of single pile with SFv= -2 at failure combination: ( )

74
 N (kN) M (kNm)
-3000 -2000 -1000 0 1000 2000 -500 0 500 1000 1500 2000 2500 3000
0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10

-12 -12
SFv=2 FSv=2

SFv=1.25 FSv=1.25
-14 -14 FSv=-2
SFv=-2
FSv=○○
SFv=○○
-16 -16

Figure 4.26. Axial loading, for different safety Figure 4.27. Moment capacity under horizontal
factors. Pile embedded in clay with Su=50kPa loading, for different safety factors

N (kN) M (kNm)
-3000 -2000 -1000 0 1000 2000 -500 0 500 1000 1500 2000 2500 3000
0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10

-12 FSv=2 -12

FSv=2
FSv=1.25
FSv=1.25
FSv=-2
-14 -14 FSv=-2
FSv=○○
FSv=○○

-16 -16

Figure 4.28. Axial loading, for different safety Figure 4.29. Moment capacity under
factors. Pile embedded in clay with Su=50kPa combination of horizontal loading and bending
moment (√ ⁄ ), for different safety

factors

75
 N (kN) M (kNm)
-10000 -8000 -6000 -4000 -2000 0 -1000 0 1000 2000 3000 4000
0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10

-12 -12 Cu=50kPa

Cu=50kPa

Cu=150kPa
-14 Cu=150kPa -14

-16 -16

Figure 4.30. Axial loading, for pile embedded in Figure 4.31. Moment capacity under horizontal
clay with a)Su=50kPa b)Su=150kPa. The factor loading, for varying undrained shear strength
of safety is SFv=1.25
Axial (MNm)

31

26

21

16

11

1
3000 2000 1000 0 1000 2000 3000
-4
Moment (kNm)
-9

Figure 4.32. The yellow area illustrates the range of influence of the axial force to the pile
moment capacity, for different factors of safety, as shown in figures 4.26 & 4.28. It is shown
that for a pile embedded in soft clay (Su=50kPa), no major change in pile bearing capacity is
noticed

76
 Chapter 5

Application to Pile Group

77
5 Application to Pile Group

5.1 Limit Equilibrium Approach

The case of a fixed‐head 3x3 Pile-Group embedded in clay with constant

undrained shear strength Su is studied. According to the 2x2 mechanisms of
failure, the equations for the 3x3 Pile-group are applied to find the failure load.

( )

where is the moment capacity of pile with zero axial loading, is the axial
( )
pile capacity to: { and is the length of the pile
( )

5.2 Finite Element Verification

The problem studied is a 3x3 Pile-group subjected to combined vertical load N,

horizontal load Q and overturning moment. The problem is analyzed as
previously with the use of the advanced Finite Element code Plaxis 3D. Figure 5.1
shows the Finite Element Model. The size of the finite element mesh is 29x29x22
taking into consideration the effect of boundaries on the pile-group’s ultimate
response and the computational time. The piles have a diameter D=1 m and the
distance between all the pile centers is 3 meters. A 9x9m plate is chosen as pile-
cap, while a sensitivity analysis is performed to ensure the fixed pile/pile-cap
connection, setting its elastic modulus equal to E=300*10^6 kPa and its
thickness of 10 meter. The piles have the aforementioned properties, i.e. an
elastic modulus of E=30*10^6, a specific weight γ=0.1kN/m3, Poisson’s ratio
v=0.2 and its behavior is governed by the Mohr-Coulomb model with c = 15262
kPa, φ=0 and tension cut-off strength equal with 7534 kPa. Interfaces are placed
between the piles and the soil enabling gapping and slippage with a friction
coefficient R=1. The soil is Clay with γ=20kN/m3 constant with the depth Su=50
kPa and Es=25000 kPa, obeying the Mohr-Coulomb model too and the Poisson’s
ratio is v=0.45 to simulate undrained conditions. The final model consists of 175000
elements with a finer discretization around the pile-group. The steps followed in the
numerical experiments are similar to the previous investigation. The total failure
Envelopes of the soil-foundation system are extracted by applying various horizontal
load – moment combinations in the normal (θ=0°) and diagonal direction (θ=45°) for
seven different safety factors from ultimate axial tensile to ultimate compressive
capacity

78
5.3 Results

The application to the 3x3 Pile-group shows major differences depending to the
undrained shear strength of the soil. Specifically, for the Pile-group embedded in soft
clay with Su=50kPa, no major differences in moment capacity of the piles are
observed (Fig. 5.10 and 5.12). However, in stiffer clay with Su=150kPa, where the
piles reach close to the failure of their axial force capacity, wide divergence to the
moment capacity of the piles in same row are noticed. (Fig. 5.19 and 5.21).
Regarding the right pile, it reaches its maximum moment capacity as Fig. 5.22 shows,
whereas the left pile, due to its tensile strength, develops almost zero moment.

79
80
Figures

81
Figure 5.1. The finite element model of 3x3 Pile-group, in clay

9m

Φ=100
3m

θ = 45°
9m

3m

Figure 5.2. Pile cap dimensions

82
 1.2

0.8

M / My 0.4 IΙ
0
0 0.2 0.4 0.6 0.8

-0.4 II
IΙ
-0.8
III
IΙI

-1.2 Q / Qy
Figure 5.3. Failure envelope of 2x2 Pile-group. The different regions define different
failure mechanisms.

I Two plastic hinges II Four Plastic Hinges at I Two Cap plastic hinges
below cap and bearing and bearing capacity
cap and below
capacity failure failure
Figure 5.4. Illustration of the failure mechanisms for combinations of Q-M
magnitudes in separate regions of the failure envelope.

83
Figure 5.5. The Pile-group mesh under horizontal loading and SFv=2

Figure 5.6. Pile-group deformed mesh Figure 5.7. Pile-group deviatoric strains
(Δγs), which indicate the regions of the
plastic hinges (2 in every pile)

84
Figure 5.8. Pile-group plastic points i.e. Figure 5.9. Pile-group vertical
the plasticized Mohr – Coulomb points deformations (uz)
(red) and the tension cut-off points
(white)

M (kNm)
-4000 -3000 -2000 -1000 0 1000 2000 3000
0

-2

-4

-6

7 8 9
-8

4 5 6
-10
1 2 3

-12

-14

-16

Figure 5.10. Moment diagram of piles in Pile-group, embedded in clay with Su=50kPa
and SFv=2, under horizontal loading

85
 N (kN)
-3500 -3000 -2500 -2000 -1500 -1000 -500 0 500
0

-2

-4

-6

-8

-10

7 8 9
-12

4 5 6
-14

1 2 3
-16

Figure 5.11. Axial force of every pile. Pile-group embedded in clay with Su=50kPa

M (kNm)
-4000 -3000 -2000 -1000 0 1000 2000 3000
0

-2

-4
SF = -2

SF = 0
-6

SF = 2
-8

-10
7 8 9

-12
4 5 6

-14
1 2 3

-16

Figure 5.12. Moment diagram of the three piles in middle row under different vertical
loads (vertical factors of safety) of pile group embedded in clay with Su=50kPa

86
Figure 5.13. Pile-group cluster with SFv=5, embedded in stiff clay (Su=150kPa)
subjected to bending moment

Figure 5.14. Pile-group deformed mesh Figure 5.15. Pile-group vertical

deformations (uz)

87
 Figure 5.16. Pile-group deviatoric strains Figure 5.17. Pile-group plastic points i.e.
(Δγs), which indicate the regions of the the plasticized Mohr – Coulomb points
plastic hinges (1 at every top of pile) (red) and the tension cut-off points
(white)

N (kN) M (kNm)
-15000 -10000 -5000 0 5000 10000 -1000 -500 0 500 1000 1500 2000 2500 3000 3500
0 0

-2 -2

-4 -4

-6 -6
4 4HALF

-8 -8 5 5HALF

6 6HALF
-10 -10
4 4HALF

-12 5 5HALF -12

7 8 9 7 8 9

-14 6 6HALF 4 5 6
-14 4 5 6

1 2 3 1 2 3
-16 -16

Figure 5.18. Axial force of the middle row of Pile- Figure 5.19. Bending moment of the middle row of
group at half of failure loading and at failure Pile-group at half of failure loading and at failure

88
 N (kN)
-15000 -10000 -5000 0 5000 10000
0

-2

-4

-6

-8

-10

7 8 9
-12
4 5 6
-14

1 2 3
-16

Figure 5.20. Axial force of Pile-group middle row, embedded in clay with Su=150kPa
and with SFv=5

M (kNm)
-1000 -500 0 500 1000 1500 2000 2500 3000 3500
0

-2

-4

-6

7 8 9
-8

4 5 6
-10

1 2 3
-12

-14

-16

Figure 5.21. Bending moment of Pile-group middle row embedded in clay with
Su=150kPa and with SFv=5. It is shown clearly the effect the different axial force to
every pile has, to the moment distribution of the pile. Regarding the compressed right
pile, the axial force reacts beneficial on the moment capacity, whereas, the axial
force of the tensioned left pile has detrimental influence to the pile moment capacity

89
 Axial (MNm)
31

26

21

16

11

1
3000 2000 1000 0 1000 2000 3000
-4
Moment (kNm)
-9

Figure 5.22. The above-mentioned results of figure 3.21 are validated by this figure,
where can be seen clearly the effect of the axial force to the moment capacity. The
yellow area illustrates the range of the axial force of the Pile-group middle row,
embedded in clay with Su=150kPa. Major changes are observed to the pile moment
capacity according to the axial force of the pile.

M (MNm)

-2 -1 0 1 2 -2 -1 0 1 2
0 0

5 5

z:m
10 10

15 δu 15
0.5 δu
0.1 δu
20 20
Pile non-linearity Soil – Pile
non-linearity

Figure 5.23. As proposed to N. Gerolymos Ph.D. thesis, the two plastic hinges of
fixed-head pile tend to approach one another near failure. This explains why the piles
of the Pile-group of Fig. 5.16 have only one plastic hinge. The plastic hinges below
cap tend to the pile head and at failure both plastic hinges coincide at pile head

90
Chapter 6

Conclusions

91
5 Conclusions

This thesis dealt with the development of a new algorithm which calculates the pile
response in terms of internal forces, under arbitrary loading. The implementation of
the method was made to single pile as well as to pile-group loading, of nine piles in
rectangular shape joined together with a pile cap. Both of which were subjected to
static monotonic loading. The constitutive model of piles that is used is the
Macroscopic Mohr-Coulomb based approach and the surrounding soil is undrained
clay.

Our first aim was to develop a new approach able to simulate in finite element
modeling the pile behavior accurately. Taking into consideration the drawbacks and
weaknesses of the existing methods a new tool is designed capable of capturing the
pile material properties and response in elasticity, in plasticity and incorporating the
effects of random loading.

Our second aim was to develop a new method of extracting the structural forces of
the pile that overcomes the drawbacks of the existing approaches and implement it
to various load cases.

Our third aim was to provide insight to the failure envelopes of a single flexible pile
and a 3x3 pile-group under combined M-Q-N loading, including the effects of soil-
pile nonlinearities. The above-mentioned method was examined and compared with
the existing approaches.

The important conclusions that were drawn from this thesis are presented below:

 A new macroscopic approach is developed under the Mohr-Coulomb i.e.

Tresca failure criterion for simulating circular piles behavior. Its verification was
focused in various concrete pile diameters (0.8m, 1m, 1.5m) with different
reinforcements (1%, 1.5%, 2%), analytically and in finite element modelling. The
results where more than satisfactory, as the approach is able to simulate the pile
behavior in elastoplasticity, in terms of moment-curvature and to include the
interaction between the external loads and the pile material properties, as well as
the soil-pile interaction. The calibration of the approach can be easily be conducted
through the following steps:

1. Extraction of the failure envelope of the studied circular pile.

2. Using the derived mathematical expressions and an optimization tool

the calibration of the parameters of the model can be performed.

92
 3. The variables are inserted in the finite element model in the
appropriate soil model together with the other elastic pile parameters

4. Verification of the approach can be done by extracting the failure

envelopes in the finite element program and comparing the moment-
curvature diagrams.

 The new method of extracting pile forces is developed.The presentation of

the pile response in terms of structural forces is an achievement of the beam
theory and expresses the pile behaviour in a macroscopic level. However, this
approach is subjected to the assumptions and simplifications of the beam
theory (e.g. Bernoulli – Euler beam theory or Timoshenko beam theory). In
view of this, the need of a tool is aroused, which incorporates the soil and
pile behaviour and their interaction at an almost microscopic level and
represents them in the same macroscopic terms of the beam theory: Axial
and Shear force and Bending moment. On the contrary to any other existing
method, this one takes into account the full pile geometry and every type of
non-linearity of the soil – structure interaction i.e. the pile plastification, the
soil non-linearity and non-linearities regarding the geometry of the model
due to sliding and gapping. The pile is not treated as one-dimensional beam
element and its response is not governed by the principals of beam theory.
Instead, it stems from the continuous medium mechanics theory, because it
manipulates the stresses that are extracted from the finite element
simulation. This means that it is possible to take into account the interaction
between the pile internal forces and the extent to which it influences the
whole pile behaviour. Additionally, since the model takes into account the
whole 3-D geometry of the soil – pile system, the pile-to-pile interaction is
considered to its full extend. Hence, the actual pile behaviour is presented in
a well-handled form and can be used for the design of the pile and its
reinforcement.

 The failure envelopes of a single flexible pile in cohesive soil are derived. The
pile is modeled by the previously derived approach. Various moment- lateral
load combinations are applied under different safety factors against axial
failure and the interaction diagrams are created. The correlation between the
axial force and the bending capacity is examined among piles with same
factor of safety and varying axial loading capacity.

 The interaction diagrams in 1x2 and 2x2 pile groups in cohesive soil are
derived. From these two we make justifiable assumptions for the 3x3 pile-

93
 group and examine its behaviour under different points of the failure
envelope. First, the pile-group is examined in soft clay, where it cannot
develop its axial capacity and minor differences between the stretched and
the compressed pile response are noticed. On the contrary, the embedded
pile-group in stiff clay, under the moment of failure, reveals a major
difference to the stretched and compressed pile bearing capacity.

 The implementation of the new method gives an opportunity to examine the

validity of the proposed Mohr-Coulomb failure criterion among varying types
of loading and the extent to which, both of them correspond to each other
and to reality as well.

94
Chapter 7
References

95
American Petroleum Institute (API), 2000, Recommended practice for planning,
designing and constructing fixed offshore platforms‐working stress design, API
Recommended Practice 2A‐WSD (RP 2A‐WSD), 21st edn.

Broms, B. 1964. Lateral resistance of piles in cohesive soils. Journal of Soil Mechanics
and Foundation Division, ASCE, 90(3): 27-63.

Comodromos EM and Papadopoulou MC (2012) "Response evaluation of horizontally

loaded pile groups in clayey soils", Geotechnique, Vol. 62, No. 4, pp 329-39.

Correia A., Pecker A., Kramer S., Pinho R. 2012. A pile-head macro-element approach
to seismic design of extended pile-shaft-supported bridges. 2nd International
Conference on Performance-Based Design in Earthquake Geotechnical Engineering,
Taormina, Italy, 28-30 May.

Crémer C, Pecker A, Davenne L. Cyclic macro-element for soil-structure interaction:

material and geometrical non linearities. International Journal for Numerical and
Analytical Methods in Geomechanics 2001; 25: 1257 – 1284.

Esmaeily A. 2001. USC-RC. Software for analyzing behavior of a single reinforced

concrete member. Version 1.0.2, USC Civil Engineering Department.

Gerolymos N. & Gazetas G., Winkler model for lateral response of rigid caisson
foundations in linear soil, Soil Dynamics and Earthquake Engineering, Volume 26,
Issue 5, May 2006, Pages 347-361

Gerolymos N. & Gazetas G. 2005. Phenomenological model applied to inelastic

response of soil–pile interaction systems. Soils and Foundations, 45(4): 119-132.

Gerolymos N., Papakyriakopoulos O., R. Brinkgreve 2014. Macroelement modeling of

piles in cohesive soil subjected to combined lateral and axial loading

Giannakou, A., Gerolymos, N., Gazetas, G., Tazoh, T., and Anastasopoulos,
I. (2010). ”Seismic Behavior of Batter Piles: Elastic Response.” J. Geotech.
Geoenviron. Eng., 136(9), 1187–1199.

Hansen, J.B. (1970). "A revised and extended formula for bearing capacity." Danish
Geotechnical Institute, Copenhagen, Denmark, Bulletin 28, pp. 5-11.

96
Harada T, Kubo K, Katayama T. Dynamic soil–structure interaction analysis by
continuum formulation method. Report of the Institute of Industrial Science, vol. 29.
The University of Tokyo; 1981 [issue no. 5].

Hill, R. 1950. The mathematical theory of plasticity. Clarenden Press, Oxford.

J.H. Atkinson (2007): “The Mechanics of Soils And Foundations”

Martin, C. M., Houlsby, G. T. 2000. Combined loading of spudcan foundations on

clay: laboratory tests, Geotechnique, 50(4): 325-338.

Matlock, H. 1970. Correlations for design of laterally loaded piles in soft clay. Proc.,
2nd Annual Offshore Technology Conference, Houston, Texas. 577-594.

Papakyriakopoulos O., Gerolymos N. (2013): Macroelement modeling of the non-

linear response of piles and pile-groups subjected to combined lateral and axial
loading, Diploma Thesis, NTUA

PLAXIS3D, 2012. Material Models Manual. Delft, The Netherlands: Plaxis b.v..

PLAXIS3D, 2012. Reference Manual. Delft, The Netherlands: Plaxis b.v..

PLAXIS3D, 2012. Scientific Manual. Delft, The Netherlands: Plaxis b.v..

PLAXIS3D, 2012. Tutorial. Delft, The Netherlands: Plaxis b.v..

Poulos H.G., Davis E.H., 1980, Pile Foundation Analysis and Design

Prager, W. (1959). An introduction to plasticity. Reading, MA: Addison-Wesley.

Randolph M.F. & Houlsby G.T. 1984. The limiting pressure on a circular pile loaded
laterally in cohesive soil. Geotechnique, 34(4): 613-623.

Randolph, M. F. (1981): The response of flexible piles to lateral loading,

Geotechnique, 31(2), 247-259

Reese LC, Cox WR, Koop FD (1974) Analysis of laterally loaded piles in sand. Proc.,
6th annual offshore technology conference, Houston, Texas, pp 473–485

Roscoe KH, Schofield AN. The stability of short pier foundations on sand. British
Welding Journal 1956; August: 343–354.

97
Roscoe, K.H. & Schofield A. N. 1956. The stability of short pier foundations on sand.
Discussion. British Welding Journal, January: 12-18.

Schnabel PB, Lysmer J, Seed HB (1972) SHAKE—a computer program for earthquake
response analysis of horizontally layered sites. EERC 72-12, Univ. of Calif., Berkeley

Sergey Edward Lyshevski, (2003): Rochester Institute of Technology Engineering and

Scientific Computations Using MATLAB, John Wiley & Sons

Skempton, A.W. 1951. The bearing capacity of clays. In Proceedings of Building

Research Congress, London. Vol. 1, pp. 180- 189

Tomilson M. J, Woodward J. (2007): “Pile Design and Construction Practice”

Varun, 2006. A simplified model for lateral response of caisson foundations, s.l.: s.n.

Walter Gander, Jiri Hrebicek (2004): Solving Problems in Scientific Computing Using
Maple and MATLAB, Springer

98
APPENDIX

Slice method

clear
format long
P=xlsread('Plaxis_Mat.xlsx',3,'A3:E200');
P(:,3)=[];
CO_plaxis=unique(P(:,1:2),'rows','sorted');
last_plaxis = length(CO_plaxis);
X_plaxis=CO_plaxis(:,1);
Y_plaxis=CO_plaxis(:,2);
sigma_plaxis=zeros(size(X_plaxis));
shear_plaxis=zeros(size(X_plaxis));
for i=1:last_plaxis
k=0;
sum1=0;
sum2=0;
for j=1:length(P)
if and(P(j,1)==CO_plaxis(i,1),P(j,2)==CO_plaxis(i,2))
k=k+1;
sum1=sum1+P(j,3);
sum2=sum2+P(j,4);
end
end
sigma_plaxis(i,1)=sum1/k;
shear_plaxis(i,1)=sum2/k;
end
%
Xmax=round(max(P(:,1))*10)/10;
Xmin=round(min(P(:,1))*10)/10;
Ymax=round(max(P(:,2))*10)/10;
Ymin=round(min(P(:,2))*10)/10;
rad=linspace(0,1.999*pi,80)';
extsize=length(rad);
bounds=[((Xmax+Xmin)/2)+((Xmax-Xmin)/2+0.005)*cos(rad),
((Ymax+Ymin)/2)+((Ymax-Ymin)/2+0.005)*sin(rad)];
%
r=linspace(0,(Xmax-Xmin)/2,25);
theta=linspace(0,2*pi,80);
[r,theta]=meshgrid(r,theta);
x=(Xmax+Xmin)/2+r.*cos(theta);
y=(Ymax+Ymin)/2+r.*sin(theta);
F_sigma=scatteredInterpolant(X_plaxis,Y_plaxis,sigma_plaxis,'linear',
'nearest');
sigma=F_sigma(x,y);
%
F_shear=scatteredInterpolant(X_plaxis,Y_plaxis,shear_plaxis,'linear',
'nearest');
shear=F_shear(x,y);
%
x=reshape(x,[],1);
y=reshape(y,[],1);
sigma=reshape(sigma,[],1);
shear=reshape(shear,[],1);
Ptemp=[x,y];
[COtemp,ia,ic]=unique(Ptemp,'rows','sorted');
sigmatemp=zeros(size(ia));
for i=1:length(ia)

99
 sigmatemp(i)=sigma(ia(i));
end
sigma=[zeros(extsize,1); sigmatemp];
sheartemp=zeros(size(ia));
for i=1:length(ia)
sheartemp(i)=shear(ia(i));
end
shear=[zeros(extsize,1); sheartemp];
CO=[bounds;COtemp];
last = length(CO);
X=CO(:,1);
Y=CO(:,2);
[v,c]=voronoin([X(:) Y(:)]);
figure;
voronoi(X,Y)
plabels = arrayfun(@(n) {sprintf('P%d', n)}, (1:last)');
Hpl = text(X, Y, plabels, 'FontWeight', ...
'bold', 'HorizontalAlignment','center', ...
'BackgroundColor', 'none');
dA=zeros(last-extsize,1);
for i=(extsize+1):last
dA(i)=polyarea(v(c{i},1),v(c{i},2));
end
A=sum(dA);
title(['Area = ' num2str(A)],'fontsize',14);
set(gca,'fontsize',14);
set(gcf, 'Position', get(0,'Screensize'));
dN=zeros(last-extsize,1);
for i=(extsize+1):last
dN(i)=dA(i)*sigma(i);
end
N=sum(dN);
%
coeff=polyfit(X((extsize+1):last),sigma((extsize+1):last),1);
%To X_plaxis den exei extsize
x0=coeff(1,2)/(-coeff(1,1));
dM=zeros(last-extsize,1);
%
for i=(extsize+1):last
dM(i)=dA(i)*sigma(i)*((Xmax+Xmin)/2-X(i));
end
M=sum(dM);

dQ=zeros(size(X_plaxis));
dQ((extsize+1):last)=dA((extsize+1):last).*shear((extsize+1):last);
Q=sum(dQ);
NMQ=[N;M;Q]

100
Volume Method

clear all
format long
P=xlsread('Plaxis_Mat.xlsx',5,'D2:H6437');
COtemp=P(:,1:3);
sigmatemp=P(:,4);
sheartemp=P(:,5);
[CO_plaxis,ia,ic]=unique(COtemp,'rows','sorted');
%last_plaxis = length(CO_plaxis);
sigma_plaxis=zeros(size(ia));
shear_plaxis=zeros(size(ia));
for i=1:length(ia)
sigma_plaxis(i)=sigmatemp(ia(i));
shear_plaxis(i)=sheartemp(ia(i));
end
X_plaxis=CO_plaxis(:,1);
Y_plaxis=CO_plaxis(:,2);
Z_plaxis=CO_plaxis(:,3);

Xmax=round(max(X_plaxis)*10)/10;
Xmin=round(min(X_plaxis)*10)/10;
Ymax=round(max(Y_plaxis)*10)/10;
Ymin=round(min(Y_plaxis)*10)/10;
Zmax=round(max(Z_plaxis));
Zmin=round(min(Z_plaxis));
rad=linspace(0,1.999*pi,80)';
extsize=length(rad);
bounds=[((Xmax+Xmin)/2)+((Xmax-Xmin)/2+0.005)*cos(rad),
((Ymax+Ymin)/2)+((Ymax-Ymin)/2+0.005)*sin(rad)];

r=linspace(0,(Xmax-Xmin)/2,25);
theta=linspace(0,2*pi,80);
zeta=linspace(Zmax,Zmin,abs(Zmin-Zmax)+1)';
zeta1=zeta;
[r,theta,zeta]=meshgrid(r,theta,zeta);
xgrid=(Xmax+Xmin)/2+r.*cos(theta);
ygrid=(Ymax+Ymin)/2+r.*sin(theta);

F_sigma =
scatteredInterpolant(CO_plaxis,sigma_plaxis,'linear','nearest');
sigmagrid = F_sigma(xgrid,ygrid,zeta);
F_shear =
scatteredInterpolant(CO_plaxis,shear_plaxis,'linear','nearest');
sheargrid = F_shear(xgrid,ygrid,zeta);
NMQ=zeros(3,size(zeta,3));
for I=1:size(zeta,3)
xloop=reshape(xgrid(:,:,I),[],1);
yloop=reshape(ygrid(:,:,I),[],1);
sigmaloop=reshape(sigmagrid(:,:,I),[],1);
shearloop=reshape(sheargrid(:,:,I),[],1);
Ploop=[xloop,yloop];
[COloop,ia,ic]=unique(Ploop,'rows','sorted');

101
 sigmalooptemp=zeros(size(ia));
shearlooptemp=zeros(size(ia));
for i=1:length(ia)
sigmalooptemp(i)=sigmaloop(ia(i));
shearlooptemp(i)=shearloop(ia(i));
end
sigma=[zeros(extsize,1); sigmalooptemp];
shear=[zeros(extsize,1); shearlooptemp];
CO=[bounds;COloop];
last = length(CO);
X=CO(:,1);
Y=CO(:,2);
[v,c]=voronoin([X(:) Y(:)]);
%figure;
%voronoi(X,Y)
plabels = arrayfun(@(n) {sprintf('P%d', n)}, (1:last)');
Hpl = text(X, Y, plabels, 'FontWeight', ...
'bold', 'HorizontalAlignment','center', ...
'BackgroundColor', 'none');
dA=zeros(last-extsize,1);
for i=(extsize+1):last
dA(i)=polyarea(v(c{i},1),v(c{i},2));
end
A=sum(dA);
title(['Area = ' num2str(A)]);
dN=zeros(last-extsize,1);
for i=(extsize+1):last
dN(i)=dA(i)*sigma(i);
end
N=sum(dN);
coeff=polyfit(X(extsize+1:last),sigma(extsize+1:last),1);
x0=coeff(1,2)/(-coeff(1,1));
dM=zeros(last-extsize,1);

for i=(extsize+1):last
dM(i)=dA(i)*sigma(i)*((Xmax+Xmin)/2-X(i));
end
M=sum(dM);
dQ=zeros(last-extsize,1);

dQ((extsize+1):last)=dA((extsize+1):last).*shear((extsize+1):last);
Q=sum(dQ);
NMQ(:,I)=[N;M;Q];
end
%Visualizations
subplot(1,3,1);
plot(NMQ(1,:),zeta1,'-b*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Axial Force (N)', 'FontSize', 20);
xlabel('(kN)','FontSize',16);
ylabel('Depth (m)','FontSize',16);
set(gca,'XAxisLocation','top','YAxisLocation','left','fontsize',14);
hold off
subplot(1,3,2);
plot(NMQ(2,:),zeta1,'-r*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Bending Moment (M)', 'FontSize', 20);
xlabel('(kNm)','FontSize',16);
ylabel('Depth (m)','FontSize',16);

102
set(gca,'XAxisLocation','top','YAxisLocation','left','fontsize',14);
hold off
subplot(1,3,3);
plot(NMQ(3,:),zeta1,'-g*','LineWidth',2,'MarkerSize',8);
hold on
plot([0 0],[Zmax Zmin], 'k:');
title('Shear Force (Q)', 'FontSize', 20);
xlabel('(kN)','FontSize',16);
ylabel('Depth (m)','FontSize',16);
set(gca,'XAxisLocation','top','YAxisLocation','left','fontsize',14);
set(gcf,'Color',[0.8,0.8,0.8],'Toolbar','none')
set(gcf, 'Position', get(0,'Screensize'));
hold off

%Stress distribution
[x_rec, y_rec, z_rec] =
meshgrid(linspace(Xmin,Xmax,100),linspace(Ymin,Ymax,100),linspace(Zma
x,Zmin,abs(Zmin-Zmax)/2+1));
sigma_rec = griddata(xgrid,ygrid,zeta,sigmagrid,x_rec,y_rec,z_rec);
shear_rec = griddata(xgrid,ygrid,zeta,sheargrid,x_rec,y_rec,z_rec);

figure;
slice(x_rec,y_rec,z_rec,sigma_rec,[],[],[Zmax:-1:Zmin]);
axis ([Xmin Xmax Ymin Ymax Zmin Zmax]);
axis equal
title('Normal Stress (kPa)');
set(gcf,'Color',[0.4,0.4,0.4])
colorbar
zoom (3)
pan on
set(gcf, 'Position', get(0,'Screensize'));
figure;
slice(x_rec,y_rec,z_rec,shear_rec,[],[],[Zmax:-1:Zmin]);
axis ([Xmin Xmax Ymin Ymax Zmin Zmax]);
axis equal
title('Shear Stress (kPa)');
set(gcf,'Color',[0.4,0.4,0.4])
colorbar
zoom (3)
pan on
set(gcf, 'Position', get(0,'Screensize'));

103