Experiences on Dike Foundations and Land Fills

on Very Soft Soils

H.-G. Kempfert, M. Raithel
Institute of Geotechnique, University of Kassel, Germany
Kempfert + Partner Geotechnik, Kassel, Wrzburg, Germany

ABSTRACT: This paper presents the land reclamation at the Elbe River in Hamburg, Germany for the necessary areaextension of the airplane dockyard (EADS) in Hamburg-Finkenwerder. The airplane dockyard (EADS) in HamburgFinkenwerder will be enlarged by approx. 140 ha (346-acres) for new branches of production, in particular for the
production of the new Airbus A 380. The area extension is carried out by enclosing the polder with a 2.4 km long dike
on very soft sludge. The necessary dike foundations were realized by about 60000 Geotextile-Encased Columns (GEC)
with a diameter of 80 cm, which were sunk to the bearing layers with depth between 4 and 14 m below the base of the
dike footing. After the enclosure of the polder the fist sand layers (until 3.0 m over sea level) filled up in the area under
buoyancy.

INTRODUCTION
The factory site of the airplane dockyard (EADS) in
Hamburg-Finkenwerder will be enlarged by approx. 140
ha for new branches of production, in particular for the
production of the new Airbus A 380.
The necessary area-extension is located in the
Mhlenberger Loch adjacent to the west of the existing
factory site. The area extension is carried out by enclosing
the polder with a 2.4 km long dike. The situation is shown
in figure 1.

section I
section II
section III
section IV

area extension

section V
section VI
section VII

new dike

Figure 1. Concept to reclaim land by the creation of a polder

A temporary enclosure is necessary, because it is only

possible to fill up the fist sand layers (until 3.0 m over sea
level) in the area under buoyancy. Without doing this it
will get stability problems, the soft soils would be
displaced into the river area. This is not allowed.
2 CONCEPT FOR LAND RECLAMATION

The original concept design in the tender documents for

enclosing the area called for a 2500 m long temporary
sheet wall to depth of a 40 m with rear-anchored raking
piles, to serve as a floodwall. Protected by the temporary
enclosure, a constant water level was to be maintained
within the area, and the first sand layers were to be filled
under buoyancy. The 346-acre area was to be raised to the
height of 5.5 m above sea level by a combination of sandtrickling, sand-sluicing and hydraulic filling. Following a
three-year consolidation of the soft soil within the
enclosure, the real flood control, a dike, was to be filled
up to the height of 9 m above sea level. In the final step,
the temporary sheet wall and the old existing dike line
were to be removed.
The value engineering concept uses geotextile encased
columns GEC as a basic foundation for the dike. After the
system is installed, the dike can be filled immediately. The
temporary sheet wall is no longer necessary and the
empoldering function will be served by the dike itself.
The necessary dike foundations were realized by about
60000 geotextile encased sand columns (System Mbius
GEC) with a diameter of 80 cm, which were sunk to the
bearing layers with depth between 4 an 14 m below the
base of the dike footing (Figure 2). The sand columns
under the base of the dike are installed by the vibro
displacement method.

Ringtrac 100 / 200

Ringtrac 100 / 200

Ringtrac 100 / 400

NN +8,85m

Comtrac 500/100 A15

Comtrac R 1000/100 A15

NN -0,6m
sludge
clay
peat
NN -8,6m

clay
sand
0,80 m

2,951 m

0,80 m

side

area ratio 10%

1,968 m

2,951 m

middle

side

area ratio 15%

area ratio 10%

Figure 2. Cross section of the value engineering concept

the horizontal support of a not encased column is not

guaranteed. Removal of the contaminated sludge would be
expensive and is in any case not permitted.
Figure 3 shows the undrained shear profile in the soft
soil and one typical ground composition in this project in
section VI (oedometric modulus for a stress level ref =
100 kN/m).

Ansatz cu1

Schlick

13 / 3

' / c'

0.0

1,0
1.0

20 / 0

CLAY
w = 77192%
= 20; c = 0 kN/m
-4.30
/ = 14/4 kN/m
Klei, jung
Eoed = 450 14
kN/m
/ 4 20

2 0 .0

SLUDGE
w = 58233% '
= 20; c = 0 kN/m
/ = 13/3 kN/m
Eoed = 500 kN/m

1 0 .0

cu [kN/m]

0,0 m NN

0 .0

In comparison with the original concept, this solution

saves a considerable amount of sand, due to both the
steeper slope (1:6 against 1:20) and a large reduction in
settlement.
In addition, with the GEC solution, it was possible to
do the foundation work and bring the dike up to the
floodproof height of 7 meters above sea level within only
eight months.
So the foundation and ground improvement system
GEC
eliminated 35000 tons of steel, since a sheet wall
was not necessary
saved 150000 m of tidal mud flat reclamation
used 1100000 m less sand to fill up the dike
(steeper slope, large settlement reductions).
produced very little noise pollution (12 vibro
displacement machines reached a noise level of
50 dba at a distance of 1000 m)
shortened construction time for the dike from 3
years to 8 months
effected a dramatic settlement reduction and a
high settlement acceleration similiar to that of
vertical drains.

2,0
2.0

3,0
3.0
4.0

/0

-6.10

5,0

5.0
6.0

PEAT
w = 75-453%
Torf, schluffig
= 20; c11
= /01kN/m20
/ = 11/1 kN/m
Eoed = 550 kN/m

7.0

/0

8.0

10,0

9.0

-9.80

Klei

3 SOIL CONDITIONS

In this area, the thickness of the soft soil layer (here

especially contaminated sludge) is between 8 to 14 m. The
reclamation site is also located in mud flats with low and
high tides twice a day.
The undrained shear strength cu in the soft soil is
between 0.4 and 10.0 kN/m.
For this reason, a conventional ground improvement
with vibro displacement piles or granular piles is not
possible, because the cu is much less than 15 kN/m and

-11.20

CLAY
20
16 / 6
w = 31-171%
= 20; c = 5 kN/m
/ = 16/6 kN/m
Eoed = 1500 kN/m

SAND
= 35;c = 0 kN/m
/ = 18/10 kN/m

Figure 3. Soil conditions (example)

/5

10.0
11.0
12.0

1) Tiefe = 0 = OK
Staffelung von
der Tiefe bleib

4 GEOTEXTILE ENCASED COLUMNS (GEC)

The foundation system Geotextile-Encased Columns

(GEC) is a further development of well-known column
foundations such as vibro displacement piles and granular
piles. In contrast to conventional column foundations,
encased columns can also be used as a ground
improvement and bearing system in very soft soils, for
example peat or sludge (undrained shear strength cu < 15
kN/m). Since 1996, the new foundation system has
proved its worth in many road and railway projects in
Germany, the Netherlands and Sweden.
To implement the GEC system, we use a regular
column grid. The diameter of both the column and the
geotextile is 0.8 m. The distance between the columns is
normally between 1.7 and 2.4 m. Based on the unit cell
concept, a single column in a virtual infinite column grid
can be considered. AC designates the column area. AE is
the influence area of a hexagonal element of a single
column in triangular grid, which can be transformed into a
circular element with an equivalent area.
0

v,c

5 DESIGN OF THE DIKE FOUNDATION

AC

On the basis of the above-described analytical calculation

model and additional FEM-calculations, the grids in table
1 were designed with more than 60000 columns using
different types of the geotextile casing Ringtrac. The
stiffness of the geotextile casing was between J = 1800
and 2800 kN/m. The maximum high tensile force of the
geotextile varied between 100 and 400 kN/m over the
cross section of the dike. The length of the columns
depended on the thickness of the soft soil along the dike
line, which varied between 4 and 14 m. For this project,
the ratio of the column area AC to the influence area AE
(AC/AE) was between 0.10 and 0.20 = 10% to 20%.

AE

v,s

h,s,tot =
h,s + h,diff

h,c

improvement method and as a bearing system for very soft

soils, because radial support is guaranteed by the
geotextile.
With a non-encased column, the horizontal support of
the soft soil must be equal to the horizontal pressure in the
column. With a geotextile-encased column, the horizontal
support of the soft soil can be much lower, due to the
radial supporting effect of the geotextile casing. This
creates a difference in horizontal stress h,diff, which
results in ring tensile forces FR in the geotextile casing.
The horizontal support depends also on the vertical
pressure over the soft soil v,s which can be much smaller.
As a result we get a stress concentration above the column
head and a lower vertical pressure over the soft soil and
therefore a large settlement reduction. Figure 4 shows the
calculation model.
On the basis of the familiar procedure for calculation
and dimensioning of gravel and sand columns, an
analytical calculation model has been developed which
takes the geotextile casing into account Raithel &
Kempfert (1999). More details are shown in Raithel
(1999) and also in Raithel & Kempfert (2000). The
derived equations can be solved by iterative process, for
which it is advisable to use a calculation program.

h,s,tot =
h,s + h,diff

Table 1. Calculation results

geotextile
ring tension
force Fr

soft soil
2 rc
2 rgeo

sand column
const. volume
(vertical drain)

2 rE

Dikesection
II
III

Figure 4. Calculation Model

IV

On the one hand, there is horizontal stress in the column

h,c due to the vertical stress v,c over the column head. On
the other hand, there is horizontal earth pressure h,tot due
to the vertical stress v,s over the soft soil as well as the
horizontal support of the casing.
As opposed to conventional column foundations,
geotextile encased columns can be used as a ground

V
VI
VII

Part

High
[m above
sea level]

Grid
AC/AE
[%]

middle
side
middle
side
middle
side
middle
side
middle
side
middle
side

+9,25
+5,50
+8,90
+5,50
+8,90
+5,50
+8,90
+5,50
+8,90
+5,50
+8,90
+5,50

17
10
15
10
15
10
15
10
15
10
20
15

ca.
number
of
columns
04.400
05.700
08.000
17.000
12.000
09.800

Settlement
[cm]
50
47
41
39
70
65
109
106
95
88
169
146

As a result of the stability calculations, we needed a

geocomposite with a high tensile strength (maximum high
tensile force 500-1000 kN/m) in the dike base
perpendicular to the dike line, to accelerate the filling of
the dike and to obtain a high degree of stability in the
initial stage of construction.
The factor ( = settlement without GEC / settlement
with GEC) of ground improvement in soft soil amounts to
about = 2.5 to 3.5. This values for the ground
improvement factors could also seen in model tests,
more details are shown in Kempfert et al (1999).
The main calculation results for the design of the dike
foundation are shown in table 1.

to better contend with tidal fluctuation (3.5 m water

level difference. At low tide, work continued with the
pontoons resting directly on the soft soil, as shown in
Figure 5. After installation, the column heads were
stabilized by filling sand between the columns. Notably,
no tidal erosion was observed.
Figure 6 shows a finished column following vibro
withdrawal of the steel pipe (open base flaps).

6 GEOTEXTILE-ENCASED COLUMN INSTALLATION

Normally, there are two installation methods in practice.

With the excavation method, an open steel pipe is driven
to the natural foundation and its contents are removed by
soil auger. The vibro displacement method, which is more
economical, is more commonly used. A steel pipe with
two base flaps (which close upon contact with the soil) is
vibrated down to the bearing layer, displacing the soft soil.
The geotextile casing Ringtrac is installed and filled
with sand. At this stage, the sand in the column is loose.
After drawing the pipe under vibration (the two base flaps
open automatically, s. Figure 6) a geotextile-encased
column filled with sand of medium density remains.
With both economy and ecology in mind, the vibro
displacement was used for the entire Hamburg project.
However, the soft soil surface along the planned dike line
varied between 0.8 above sea level to 2.5 m below sea
level. Therefore, different construction methods were
necessary to install the GEC foundation for the dike.

Figure 5. Vibro displacement method from pontoon lying on

soft soil.

The majority of the columns was installed using

equipment operating from offshore pontoons (110 11 m)

Figure 6. Installed column after drawing the steel pipe under

vibration

A further GEC construction method was used for

numerous road and railway projects in Germany, the
Netherlands and Sweden. The vibro displacement machine
stood on top of the installed columns, with mats under the
120-ton unit to facilitate load distribution. This land
construction method is shown in Figure 7.

Figure 7. The well-tested vibro displacement method on land.

The displacement of the soft soil leads to an uplifting of

the soft soil within and around the column grid. The
heaving of the soft soil produced deformations at the

after installation the columns

before installation the columns
0
1
2
3

soft soil depth in m

Therefore after 39 weeks the necessary high water safety

could be reached. In figure 9 the measured values of the
settlements in the dike section VI are shown.

High+ m NN

8.00

max. settlement [m]

surface of the grid. The lifting was measured at up to 3-8 %

of the column depth.
Liquefaction of the soft soil by compaction energy was
not observed. Measurements showed an increase in the
undrained shear strength of the soft soil surrounding the
columns. Figure 8 shows one result of the measurements
of the depth of the soft soil before and immediately after
installation of the columns. Further, an increase by a
factor of 2 in the shear strength of the surrounding soft
soil was measured, which shows the additional stabilizing
effect of the installation method.

prognosis
in situ

6.00
4.00
2.00 0

28

56

84 112 140 168 196 224 252

0.00
0.20

days

0.40
0.60

measurement
prognosis

0.80
1.00
1.20

Figure 9: Measured settlements in section VI

5
6
7
8

8 FILLING

9
10
11
12
0

10

20

30
40
cU in kN/m2

50

60

Figure 8. Increase of the undrained shear strenght in the soft soil

between the columns in comparison before and after installation
the columns

Protected by the dike, a constant water level was to be

maintained within the area, and the first sand layers were
to be filled under buoyancy.
Generally the total area was divided in different filling
sections according to the different soil conditions. In
Figure 10 the approximate division of the area is show.

7 MEASUREMENTS OF DIKE BEARING AND

DEFORMATION BEHAVIOUR

Due to the different soil conditions along the dike length 7

measurement cross sections are necessary.
In a typical measurement cross section 4 groups with
one earth pressure- and one water pressure gauge above
the soft soil layer, as well as two piezometers in the soft
soil are placed.
In addition in each cross section a horizontal and two
vertical inclinometers for the examination of the
deformation behaviour are used.
On the basis of the measurements it can be shown, that
the real soil conditions are better in opposite to the
documented soil parameters in the tender documents,
especially with regard to the consolidation behaviour.
Due to high effectiveness of the foundation system, the
dike could be constructed in a building time of approx. 9
months about 7 m height (in spite of hardly avoidable time
deficits in the initial phase).

Figure 10. Division of the area for filling of the first sand layers

A special section in the total area is the so called

Teilflche Sektionsbau, shown at the right corner of the
total area in figure 10.

The filling of this section must be started before the dike

was finished. For enclosure of this section, a small
temporary dam up to 3 m above sea level, founded on
about 4000 geotextile encased columns, was builded.
After the filling up to a high of about 3.1 m to 3.6 m
above sea level, the working level for the installation of
vertical drains was reached.
The vertical drains are inserted in a triangular grid with
grid distances of 0.5 m to 1.0 m. Figure 11 shows the
installation.

SUMMARY AND CONCLUSIONS

The factory site of the airplane dockyard (EADS) in

Hamburg-Finkenwerder will be enlarged by approx. 140
ha (346-acre). The necessary area-extension is located in
the Mhlenberger Loch adjacent to the west of the
existing factory site. The area extension is carried out by
enclosing the polder with a 2.4 km long dike.
The necessary dike foundations were realized by about
60000 geotextile encased sand columns (System Mbius
GEC) with a diameter of 80 cm, which were sunk to the
bearing layers with depth between 4 and 14 m below the
base of the dike footing. Due to the foundation system
Geotextile Encased Sand Columns (GEC) the dike could
be constructed on the subsoil with very small shear
strength and distinct deformability in a building time of
approx. 9 months to a safe high water height.
The 346-acre area was to be raised to the height of 5.5
m above sea level by a combination of sand-trickling,
sand-sluicing and hydraulic filling. Figure 13 shows the
area with filling in February 2002.

Figure 11. Installation of vertical drains in detail

settlements [cm]

fill high [m above sea level]

After the installation of the vertical drains the area was to

be raised to the minimum height of 5.5 m above sea level
by a hydraulic filling.
To minimize the settlements after the construction
time, due to the strict settlement requirements of 10 to 30
cm, a temporary surcharge by a higher filling (up to 10 m
above sea level) was used.
In figure 12 a typical measured time-settlement curve
after filling is shown.
8

Figure 13. Overall view of the area in February 2002

soil conditions

REFERENCES

2,6 m sludge
2

1,1 m sand

0
40 0

20

40

60

80

100

time [days]
80

120

140

160

120
160

180

4,9 m clay,
peat

Installation of vertical drains

200

Figure 12. Measured time-settlement curve after filling

200

Raithel, M. & Kempfert, H.-G. 1999. Bemessung von geokunststoffummantelten Sandsulen. Bautechnik 76. Heft 11: 983-991.
Raithel, M. 1999. Zum Trag- und Verformungsverhalten von geokunststoffummantelten Sandsulen. Schriftenreihe Geotechnik. Universitt
Gh Kassel. Heft 6. Kassel.
Raithel, M. & Kempfert, H.-G. 2000. Calculation Models for Dam
Foundations with Geotextile Coated Sand Columns. Proc.
International Conference on Geotechnical & Geological Engineering
GeoEng 2000. Melbourne.
Kempfert, H.-G. et al. 1999. Model Tests for Analysis of the Bearing and
Deformation Behaviour of Column Foundations. Geotechnical
Engineering for Transportation Infrastructure. Balkema, Rotterdam.