Applications of Coupled Eulerian-Lagrangian

Method to Geotechnical Problems with Large

Deformations
G. Qiu, S. Henke, and J. Grabe
Institute of Geotechnical and Construction Engineering, Hamburg University of Technology,
Germany
Abstract: In recent years, finite element method has been considered the main tool for solving
geomechanical boundary value problems. Studying geotechnical problems like collapse analysis
of a quay wall construction or pile penetration into the subsoil, it is evident that large deformation
of the soil body can occur. The main problems in these analyses are contact problems and large
deformations of the FE-mesh. Abaqus, version 6.8 now provides a new tool, the Coupled EulerianLagrangian method (CEL), to deal with large deformation problems. With this new method the
drawback of the FE-method, which originates from its reliance on a lagrangian mesh, can be
overcome. In this paper, two benchmark problems are analyzed using the CEL-method, the results
are compared to classic implicit- and explicit-solutions. Concluding, the application of CEL for
pile penetration analysis is investigated.

Keywords: Abaqus, finite element method, Coupled Eulerian-Lagrangian, large deformation

analysis, soil mechanics, geotechnics, strip footing, anchor plate, pile penetration.

1. Introduction
In recent years, finite element method has been considered the main tool for solving geotechnical
problems. A three-dimensional FE-simulation of a quay wall construction in service state at the
port of Hamburg has been done by Mardfeldt (2005). Regarding exceptional load cases like
extreme tide, scouring, or grounding of a ship the quay wall construction may reach its limit state.
In such cases large deformations of the soil and the structure can occur. Another application with
large deformations in geotechnical engineering is the analysis of pile penetration into the subsoil.
Several researchers investigated the pile penetration process using finite element method. First
investigations have been made by Mabsout and Tassoulas (1994) using a special zipper-typetechnique to allow the simulation of discrete hammer blows on a prebored pile. This technique has
been extended by Cudmani (2001) to simulate the cone penetration test using an axisymmetric
model. A comparison of different pile installation methods (pile jacking, vibratory pile driving,
impact pile driving) has been done by Mahutka et al. (2006). These investigations are also
constrained to axisymmetric calculations. Henke and Grabe (2006) extended this modeling

2009 SIMULIA Customer Conference

technique to allow three-dimensional analyses of pile installation so that the installation of piles
with open cross-section is possible (Henke and Hgel, 2007).
It is evident that finite element method has many problems solving geotechnical problems with
large deformations. Especially, contact problems and large mesh distortions may occur so that a
convergent solution often cannot be found. In this paper, simulations of large deformation
problems using Abaqus newly built-in feature CEL are fulfilled and the results are compared to
those received by classical implicit and explicit calculations to show the capabilities of this tool.

2. Numerical methods
2.1

Implicit analysis

The implicit solver is implemented in Abaqus/Standard. Because of implemented convergence

criteria the implicit solver is implicitly stable. The implicit procedure uses an automatic
incremental strategy based on the success rate of a full Newton iterative solution method. For
every step the quantities of the time step t and t+t are taken into account. This means that for
every time step a nonlinear system of equations has to be solved using e.g. Newmarks method.
2.2

Explicit analysis

The explicit solver is a numerical solution algorithm to solve the equation of motion:

Cu Ku f
Mu
Using the explicit solver the equation of motion is solved for a certain time step by direct
integration using the central difference rule. This method is especially effective, if a lumped or
diagonal mass matrix exists. In this case, the unknown deformations appear only at one side of the
system of equations so that the solution is easy to receive. The explicit solution procedure is not
implicitly stable. A stable time increment tcrit has to be defined for which the solution is
conditionally stable.

t crit

Le
cd

Le is the minimum element length and cd the dilatory wave speed. The main problem using an
explicit solver is that solution control is very difficult because no convergence criterion exists.
2.3

Coupled Eulerian-Lagrangian (CEL) analysis

If a continuum deforms or flows, the position of the small volumetric elements changes with time.
These positions can be described as functions of time in two ways:
Lagrangian description: the movement of the continuum is specified as a function of its initial
coordinates and time.

2009 SIMULIA Customer Conference

Eulerian description: the movement of the continuum is specified as a function of its

instantaneous position and time.
In simulations with Lagrangian formulation the interface between two parts is precisely defined
and tracked. In these simulations large deformation of a part leads to hopeless mesh and element
distortion. In Eulerian analysis a Eulerian reference mesh, which remains undistorted and does not
move, is needed, to trace the motion of the particles. The advantage of a Eulerian formulation is
that no element distortions occur. Disadvantageously, the interface between two parts cannot be
described as precise as if a Lagrangian formulation is used. Numerical diffusion can happen
during the simulation. A Coupled Eulerian-Lagrangian (CEL) method, which attempts to capture
the strength of the Lagrangian and the Eulerian method, is implemented in Abaqus/Explicit. For
general geotechnical problems, a Lagrangian mesh is used to discretize structures, while a
Eulerian mesh is used to discretize the subsoil. The interface between structure and subsoil can be
represented using the boundary of the Lagrangian domain. On the other hand, the Eulerian mesh,
which represents the soil that may experience large deformations, has no problems regarding mesh
and element distortions.

3. Benchmark
3.1

Strip footing problem

In this section, three different numerical solution algorithms (Implicit-, Explicit- and Coupled
Eulerian-Lagrangian Method) are compared solving a strip footing problem, which imposes severe
requirements on the computational procedure and is an often cited benchmark problem with an
analytical solution, see Fig. 1.

p
smooth

rough

4m

d=2m

von Mises Material

D=4m

Figure 1. Geometry and boundary conditions of penetrating a strip footing into

cohesive soil.

2009 SIMULIA Customer Conference

This plane strain problem has been analytically solved by Hill (1950) using the slip line theory.
According to Hill (1950) the maximum punch pressure for this problem with a ratio d/D = 0.5
(Fig. 1) can be calculated with:

p ( 2 )c ,
where c is the shear strength of the soil which is described using a von MISES material.
Numerical analyses have been executed by van Langen (1991) using a TRESCA material.
Analogue to his work in this study the von MISES material parameters in Tab. 1 are used for the
analyses.
Table 1. Material parameters for cohesive soil.
Property
Value

G
[kpa]
1000

[-]
0.49

c
[kPa]
10

The side of the footing is modeled smooth, whereas the base is rough. The footing, which
penetrates into a cohesive but weightless soil with dimension 4 m x 4 m, has a width of 2 m and a
height of 1 m. The strip footing is discretized as a rigid body. The plane strain problem is modeled
two-dimensionally using the implicit- and the explicit-solution algorithm. The subsoil is meshed
using linear elements with reduced integration (element type: CPE4R). Regarding the Coupled
Eulerian-Lagrangian analysis the penetration process must be simulated three-dimensionally.
Three-dimensional Eulerian elements are used to discretize the soil body (element type: EC3D8R).
The load-displacement curves that have been obtained out of the comparative analyses are
depicted in Fig. 2. The maximum reaction force is reached at a punch indentation of less than
0.1 m in all analyses. The agreement between numerical solutions and the analytical solution is
very satisfactory. The difference remains within 8%. After reaching a maximum reaction force the
solution of the CEL-analysis remains nearly constant, whereas the solution of implicit- and
explicit-method increases continuously.
The increasing reaction force in the implicit analysis can be explained considering Fig. 3. This
figure shows the calculated contact stresses underneath the strip footing for the three investigated
methods. Stress concentrations can be seen at the edge of the footing regarding the implicit
calculation which cannot be explained physically. According to the theoretical solution, the
contact pressure cannot be greater than ( 2 )c . The contact pressure must be evenly distributed
at the bottom of the footing. There should be no stress concentration at the corner of the footing.
In contrast, the stress concentration at the corner regarding the implicit analysis increases with
growing punch indentation. As shown in Fig. 4, the velocity gradient near the edge of the footing
is very high. The soil is pushed down, slips sideways and then moves upwards. The velocity field
is not uniquely to define. This point is well known as singular plasticity point. The velocity
gradient is too high to be simulated using the implicit method. The soil nearby the corner can only
move down and than moves sideways. An upward motion cannot be found in Fig. 4. To overcome
the effect of singular plasticity points, van Langen (1991) incorporated some potential slip lines
into his implicit model. Regarding the explicit analysis the element at the edge of the footing is
distorted extremely. Furthermore, an upward motion cannot be found, too. Due to the observed
4

2009 SIMULIA Customer Conference

distorted elements no stress concentration at the edge of the footing occurs. CEL-method can
overcome the restrictions of the purely lagrangian analyses. The soil body can be deformed freely.
normalized pundch pressure p/c [-]

7
6
5
4
3
2
Exact Solution (2+)
Implicit
Explicit
CEL

1
0
0

0.1

0.2

0.3

0.4

0.5

punch indentation [m]

Figure 2. Load displacement curve for penetration of a strip footing into a

cohesive soil.

1
2
3
4
5

Implicit
Explicit
CEL

strip footing
0

normalized pundch pressure p/(2+)c

normalized pundch pressure p/(2+)c

strip footing
0

1
2
3
4
5

Implicit
Explicit
CEL

Figure 3. Computed contact pressure underneath the strip footing at collapse (left)
and punch indentation of 0.5 m (right) for the three investigated methods.

2009 SIMULIA Customer Conference

Implicit

Explicit

CEL

Figure 4. Velocity field of the strip footing problem after a punch indentation of
0.5 m for the three investigated numerical methods .
3.2

Anchor plate problem

In this section, the three investigated numerical methods are used to solve a circular anchor plate
problem. A smooth circular anchor plate, with a radius of 1 m and a thickness of 0.1 m, is pulled
out of a non-cohesive soil, see Fig. 5. The soil has a friction angle = 30. The Druck-Prager
material model is used to simulate the soil. In this material model the friction angle is needed
which can be calculated as follows:

tan

6 sin .
3 sin

The used material parameters are given in table 2.

Table 2. Material parameters for non-cohesive soil.
Property
Value

E
[kPa]
35000

[-]
0.3

[]
50.2

K
[-]
1.0

[t]
2.0

The pull out process is simulated axially symmetric in the implicit- and explicit-analysis, whereas
the CEL-analysis has to be done three-dimensionally. Same as in the strip footing problem the
anchor plate is modeled as a rigid body. Three-dimensional Eulerian elements are used to
discretize the soil (element type: EC3D8R) in the CEL analysis. In the implicit- and explicit
6

2009 SIMULIA Customer Conference

analyses axisymmetric linear continuum elements with reduced integration are used to discretize
the non-cohesive soil body (element type: CAX4R).

60
2m

25 m

8m

No analytical solution seems to be available for this anchor plate problem. The maximum vertical
pull out force can approximately be calculated out of the weight of the soil body, which lies within
the slip lines above the anchor plate, see Fig. 5. This soil body in the present boundary value
problem has a weight of W = 6399 kN.

30 m

Figure 5. Geometry and boundary conditions of the investigated circular anchor

plate problem.
Due to convergence problems the implicit-analysis does not provide any solution. It is well
known, that the pull out velocity has a large influence on the uplift load. Fig. 6 illustrates the load
displacement curves for different pull out velocities. As expected, a high pull out velocity leads to
large oscillations of the result curves. Regarding the calculation time, a pull out velocity of 0.5 m/s
is judged to be slow enough for the following analyses.
Fig. 7 depicts the load displacement curves of the calculations using the explicit-solution
algorithm respectively CEL-method. The results of the two analyses match very well until the
peak value is reached. The after-peak behavior can also be explained regarding the velocity field
received out of the two simulations with the different numerical methods, see Fig. 8. This figure
depicts the velocity field of the anchor plate problem with explicit- and CEL-method for an uplift
of 1 m and 5 m. There are only small differences between the two methods regarding the velocity
field at 1 m uplift, see Fig. 8 (a) and (b). The soil above the anchor plate is lifted up within the
shear band and moves upwards. At an uplift of 5 m of the anchor plate, the soil should fall down
and fill the hole under the anchor plate. Regarding the explicit analysis this process is restricted
due to the distorted lagrangian elements. The elements at the edge of the anchor plate are pulled
upwards by the anchor plate. This prevents a horizontal motion of the nearby soil into the hole
under the anchor plate. Comparing the CEL-analysis the soil is able to move horizontally towards
the hole under the anchor plate without any restriction.

2009 SIMULIA Customer Conference

normalized uplift load F/W [-]

normalized uplift load F/W [-]

0.5 m/s
1 m/s
2 m/s

1
0.8
0.6
0.4

Explicit

0.2
0
0

2
3
uplift anchor plate [m]

0.5 m/s
1 m/s
2 m/s

1
0.8
0.6
0.4

CEL

0.2
0
0

uplift anchor plate [m]

Figure 6. Load displacement curve for pull out an anchor plate by different pull out
velocities in explicit- analysis (left) and CEL-analysis (right).

Figure 7. Load displacement curve for pull out an anchor plate from non-cohesive
soil.

4. Numerical simulation of the pile penetration process

4.1

Using finite element method to simulate pile installation

Although the use of piles as deep foundations is state-of-the-art for many decades the mechanisms
that occur during installation and the effects on the surrounding soil and adjacent structures with
respect to the installation method are not fully investigated. In recent years, finite element method

2009 SIMULIA Customer Conference

is used by many authors to simulate a part or even the whole penetration process. Mainly, there are
three different approaches:

Simulation of the pile driving process by cavity expansion, for example published by
Chopra and Dargush (1992). In these analyses a predefined hole is expanded horizontally
to simulate the installation process. The main disadvantage of this approach is that the
pile toe resistance is not modeled correctly and that it is nearly impossible to consider
dynamic effects.

Simulation of single hammer blows at a pile which is simulated prebored to a certain

depth (e.g. Mabsout and Tassoulas, 1994). In these analyses dynamic effects can be
considered but the disadvantages are that only single hammer blows can be computed and
that only a small part of the penetration process is simulated. This means that the effects
of pile driving on the surrounding soil cannot be predicted precisely.

Simulation of the whole penetration process using a zipper-type technique, see e.g.
Cudmani (2001), Mahutka et al. (2006) or Henke (2008). To allow penetration of the pile
into the continuum a rigid tube is modeled into the axis of penetration. This tube is in
frictionless contact with the surrounding soil. During the simulation, the pile slides over
the tube and contact between pile and soil can be established. This approach makes it
possible to simulate several meters of penetration using different installation methods.
But there are main problems regarding mesh distortions and the contact between pile and
soil. Furthermore, the rigid tube prevents the pile from drifting out of the vertical axis and
imposes severe problems in complex soil-structure-interaction problems.

In the following section the capabilities of CEL to simulate the pile installation process are
examined simulating the installation of a steel-pile into the soil.
4.2

Geometry

The penetration of a 10 m long Peiner steel-pile PSt 500/158 into the subsoil is investigated. An
idealized cross-section of the PSt 500/158 is shown in Fig. 9. Due to symmetry only one fourth of
the whole model is considered in the three-dimensional analysis. The pile is modeled with
conventional Lagrangian continuum elements with reduced integration (C3D8R), whereas the soil
body is meshed using Eulerian elements (EC3D8R). The initial lateral stress coefficient K0 is set to
0.5 for the sand with a friction angle = 30.

2009 SIMULIA Customer Conference

(a)

(b)

(c)

(d)

Figure 8. Velocity field for the anchor plate problem (a) 1 m uplift in explicitanalysis; (b) 1 m uplift in CEL-analysis; (c) 5 m uplift in explicit-analysis; (d) 5 m
uplift in CEL- analysis.

10

2009 SIMULIA Customer Conference

17

12

506

472

20.2

17

381

Figure 9. Idealized cross-section of a PSt 500/158.

4.3

Material model

A simple linear-elastic, perfectly plastic material model has been used to describe the material
behavior of the Peiner steel-pile. The properties of the pile correspond to those of steel: elastic
modulus E = 210 GPa, Poisson ratio = 0.3 and yield stress f = 235 MPa. The sand is modeled as
an elastic-plastic material using the modified Drucker-Prager/Cap model as it is implemented in
Abaqus. The material parameters are given in Table 3 below.
Table 3. Material parameters for soil.
Property

in
vol
0

Value

[Mpa]
20

[-]
0.333

[kPa]
0.1

[]
50.2

[-]
0.1

[-]
0.0009

[-]
0.01

[-]
1

Figure 10 depicts the cap hardening curve used in the following calculations. During pile
installation the soil experiences high compaction and many changes between loading and
unloading conditions. These are effects which cannot be fully simulated using the modified
Drucker-Prager/Cap model so that it is necessary to question the results of the numerical
calculations critically.
4.4

Contact formulation

The contact between Eulerian domain and Lagrangian domain is discretised using the general
contact algorithm, which enforces the use of the penalty contact method. The penalty contact
method is less stringent compared to the kinematic contact method. Currently, the finite-sliding
formulation is the only available tracking approach for general contact. Thus, the finite-sliding
formulation allows arbitrary motion of the surface, such as separation, large sliding and rotation, it
is well suited to simulate highly nonlinear processes with large deformation. The frictional sliding
2009 SIMULIA Customer Conference
11

at the interface between pile and soil is modeled using the Coulomb friction contact law with a
friction coefficient = 0.176.

Figure 10. Left: Contour plot of horizontal stress distribution; Right: Contour plot
of vertical stress distribution after 10 m of jacking a Peiner steel-pile into the
subsoil.
4.5

Installation process

The pile driving process is modeled displacement controlled, dynamic effects are not taken into
account. This means that the pile is jacked into the subsoil. Sudden movements at the beginning of
the installation cause stress waves, which lead to a noisy solution, see Fig. 11. In order to reduce
the noise in the present quasi-static analysis, the installation velocity should be increased as
smooth as possible. Two different pile jacking simulations are carried out. First, the pile is
installed with constant velocity (test 1) and in the second analysis a smooth increase of velocity in
the first second is considered (test 2). The penetration velocity is shown in Fig. 11. Fig. 11 also
depicts the development of pile resistance during penetration regarding the two different
simulations. As expected, the solution from test 1 is very noisy, while the results from test 2 are
quite smooth.

5. Exemplary numerical results

In this section, exemplary results of jacking a Peiner steel-pile into the subsoil are presented and
checked for feasibility. In Fig. 12 the stress distributions around the Peiner steel-pile after 10 m of
pile jacking are depicted. Regarding the horizontal stresses it becomes clear that the stresses
decrease directly at the pile shaft up to a very low value. A significant increase can only be
observed in a certain distance from the pile. Comparing this result with validated results out of
literature using classic finite element method it can be stated that these results differ from those
e.g. shown by Henke (2008) or Mahutka (2007). In these contributions, the authors show that
12

2009 SIMULIA Customer Conference

jacking a pile into the subsoil leads to increasing horizontal stresses. This difference between CEL
results and those found in literature can be explained by the used material model for the subsoil.
The used Drucker-Prager soil model is not well suited to simulate highly non-linear problems with
large deformations like pile penetration. Especially, compaction of the subsoil and dilatancy and
contractancy cannot be fully reproduced. Mahutka (2007) and Henke (2008) used a hypoplastic
constitutive law to describe the soil behavior which is better suited to simulate the soil behavior
during pile installation because it is able to reproduce typical soil characteristics like dilatancy,
contractancy, different stiffness for loading and unloading and a dependency on void ratio. So it
can be concluded that it is important for soil mechanical purposes that the implementation of highquality constitutive soil models using VUMAT must be made possible for CEL calculations.
pile resistance [kN]

1.2

velocity [m/s]

1
0.8
0.6
0.4
Test 1
Test 2

0.2
0
0

0.5

1.5

time [s]

80
70
60
50
40
30
20
10
0

Test 1
Test 2
0

0.1

0.2

0.3

0.4

0.5

penetration depth [m]

Figure 11. Left: Installation velocity; Right: Development of pile resistance during
pile installation with constant velocity (test 1) and with smooth step
amplitude (test 2).
Regarding the vertical stress distribution an increase of stress can be remarked at the pile toe. This
result is in good accordance to those found in literature. Due to pile jacking, especially the stresses
at the pile toe can reach very high values.
The deformed ground surface after 10 m of jacking the steel-pile into the soil body is depicted in
Fig. 13. It can be seen that a typical funnel around the pile has been generated due to the pile
jacking process. Soil is pulled downwards with the penetrating pile so that the funnel emerges.
This result is in good accordance with in-situ observations and numerical results using classic
finite element analyses, see Mahutka (2007) or Clough and Chameau (1980). Nevertheless, it has
to be stated that the formation of a funnel as observed depends on many factors like properties of
the surrounding soil, installation method, shape of the installed profile or friction between pile and
soil.
Concluding, it can be stated that CEL is a well suited method to simulate the pile penetration
process but especially the lack of adding high-quality constitutive laws for the soil leads to results
which are not 100% feasible. Nevertheless, it has been announced that in Abaqus, version 6.8-EF
the use of VUMAT in combination with CEL-analysis will be made possible so that it can be
presumed that this tool will be a very powerful tool to solve highly complex boundary value
problems in soil mechanics.

2009 SIMULIA Customer Conference

13

SVAVG, SVAVG11
(Avg: 75%)
+0.000e+00
7.821e+01
1.564e+02
2.346e+02
3.128e+02
3.910e+02
4.693e+02
5.475e+02
6.257e+02
7.039e+02
7.821e+02
8.603e+02
9.385e+02

SVAVG, SVAVG33
(Avg: 75%)
+0.000e+00
1.304e+02
2.608e+02
3.911e+02
5.215e+02
6.519e+02
7.823e+02
9.127e+02
1.043e+03
1.173e+03
1.304e+03
1.434e+03
1.565e+03

Y
Z

Y
Z

Figure 12. Left: Contour plot of horizontal stress distribution; Right: Contour plot
of vertical stress distribution after 10 m of jacking a Peiner steel-pile into the
subsoil.

Figure 13. Deformed ground surface after 10 m of pile jacking

14

2009 SIMULIA Customer Conference

6. Conclusion and outlook

According to two benchmark problems it can be concluded that the CEL-method is well suited to
solve geotechnical problems with large deformations of the soil body and the interacting structure.
Solving soil-structure-interaction problems singular plasticity points often can be found at the
edge of the structures. The free deformable soil in an Eulerian domain can overcome the effect of
singular plasticity points automatically. Due to large deformations of the structure a clearance
between soil and structure can occur. This leads to non-linearities at the boundary and can induce
convergence problems if an implicit analysis is chosen. Regarding the two investigated benchmark
problems it can be stated that only CEL-method is able to predict the resulting loads at limit state
correctly because the Eulerian mesh is not restricted by mesh or element distortions.
Furthermore, an example of simulating pile penetration into the subsoil is shown to show the
capabilities of CEL to simulate complex geotechnical problems involving large deformations. It
can be stated that CEL is well suited to simulate the pile penetration process without the
limitations of a classical finite element formulation.
Nevertheless, there are some features that have to be improved. Especially, it is necessary for soil
mechanical purposes that user subroutines (esp. VUMAT) are made available for Eulerian
domains. Soil is a complex material which cannot be described correctly using the Abaqus builtin constitutive models. Especially, if complex boundary value problems involving large
deformations and/or dynamic effects have to be solved, high-quality constitutive models are
necessary.
If the use of VUMAT in combination with CEL is made possible, CEL is expected to be a very
powerful tool to deal with geotechnical boundary value problems, which are nowadays hardly
solvable using finite element method.

7. References
1. Chopra, M.B., Dargush, G.F., Finite-Element Analysis of time-dependant large-deformations
Problems. International Journal for Numerical and Analytical Methods in Geomechanics 16,
101-130, 1992.
2. Clough, G.W., Chameau, J.-L., Measured Effects of Vibratory Sheetpile Driving. Journal of
Geotechnical Engineering Division, 106(10): 1081-1099, 1980.
3. Cudmani, R.O., Statische, alternierende und dynamische Penetration in nichtbindigen Bden.
Dissertation. Verffentlichungen des Institutes fr Bodenmechanik und Felsmechanik der
Universitt Fridericiana in Karlsruhe, Heft 152, 2001.
4. Henke, S., Grabe, J., Simulation of pile driving by 3-dimensional Finite-Element analysis.
Proceedings of 17th European Young Geotechnical Engineers Conference, Zagreb, pp.215233, 2006.
5. Henke, S., Hgel, H.M., Rumliche Analysen zur quasi-statischen und dynamischen
Penetration von Bauteilen in den Untergrund. Tagungsband zur 19. deutschen AbaqusBenutzerkonferenz in Baden-Baden, Artikel 2.13, 2007.
2009 SIMULIA Customer Conference
15

6. Henke, S., Herstellungseinflsse aus Pfahlrammung im Kaimauerbau. Dissertation.

Verffentlichungen des Instituts fr Geotechnik und Baubetrieb der Technischen Universitt
Hamburg-Harburg, Heft 18, 2008.
7. Hill, R., The Mathematical Theory of Plasticity, Oxford, 1950.
8. van Langen, H., Numerical analysis of soil-structure interaction , Ph.D. Thesis, Delft
University of Technology, 1991.
9. Mabsout, M.E., Tassoulas, J.L., A finite element model for the simulation of pile driving.
International Journal for Numerical and Analytical Methods in Geomechanics, 37, 257-278,
1994.
10. Mahutka, K.-P., Knig, F., Grabe, J., Numerical modeling of pile jacking, driving and vibro
driving. Proceedings of International Conference on Numerical Simulations of Construction
Processes in Geotechnical Engineering for Urban Environment (NSC06), Bochum, pp. 235246, 2006.
11. Mahutka, K.-P., Zur Verdichtung von rolligen Bden infolge dynamischer Pfahleinbringung
und durch Oberflchenrttler. Dissertation. Verffentlichungen des Instituts fr Geotechnik
und Baubetrieb der Technischen Universitt Hamburg-Harburg, Heft 15, 2007.
12. Mardfeldt, B., Verformungsverhalten von Kaimauerkonstruktionen im Gebrauchszustand,
Promotionsschrift, Verffentlichungen des Instituts fr Geotechnik und Baubetrieb der
Technische Universitt Hamburg-Harburg, Heft 11, 2005.

Acknowledgement
The authors thank the German Research Foundation (DFG) for funding the present project in the
framework of the post-graduate programme Seaports for container ships of future generations.

16

2009 SIMULIA Customer Conference