Design and Analysis of Approach Embankment Bridges

over Soft Soils

Carmo Cardoso*, ¶
Department of Civil Engineering, Instituto Superior Técnico, Universidade Técnica de Lisboa, Portugal

Abstract
The growth of urban areas near watersides induces the need of construction over compressible soils. This paper
presents a study on the behavior of approach embankments bridges over clay soils with special focus on its
foundations. In order to achieve results several projects were developed and the results were analyzed and
achieved based on the comparison between software Plaxis 2D modeling and the experimental embankments.

An equal set of base data was considered for all the projects. The underlying data is equal to the embankment
AE1 data, in order to represent the same real problem. The AE1 is one of the three experimental embankment led
to breakage, developed within the frame of the Magnani doctoral work in 2006. Four models were projected
ranging from constructive methods: from the more deformable - conventional embankment with geo-
reinforcement at the base and foundation treated with vertical drains built-in stages - to the more rigid- staked
embankment - passing on the intermediate - embankment over stone-columns and embankment over jacketed
stone-columns.

Keywords: Soft soils, Embankment, Wick drains, Stone Columns.

Introduction Presentation of the embankment AE1

In large metropolitan areas, the soil occupation The experimental embankment AE1 was settled
took place essentially in coastal areas due to the along with two others, AE2, AE3 over a soft clay
strategic location a ease of communications and deposit, located in the interior of the south bay of
transport together with the economic point of the Santa Catarina island, in the city of
view. However, these areas are surrounded by Florianópolis, State of Santa Catarina, in
soils shaped essentially by sedimentation, soils southern Brazil, presented in Fig.1.
eroded and carried by water and by alluvial
deposits of soft clay. These soils were so far
avoided due to their diminutive mechanical
characteristics, less favorable to groundwork.

The construction on soft clay soils remains a real

problem today and it is targeted in studies, given
the need to build new urban infrastructures and
Fig. 1 – Location [1]
transportation on their surface, as the urban
centers continuously grow.

¶
E-mail: [email protected] 1
Master Student
 Carmo Cardoso

The embankments were constructed on the Side edges were also planned to reduce the
scope of the doctoral thesis of Henrique Magnani three-dimensional effects of embankments due to
de Oliveira [2] . The main objective was to study its limited length, to minimize potential disruptions
the behavior of reinforced embankments and displacements in the longitudinal direction to
executed over a very soft clay deposit in a regime sections, and also to prevent the disruption of a
rapid construction and are part of the Via landfill that could influence the conditions of
Expressa Sul project, a road that links the stability given the vicinity.
downtown area to the southern part of the island
The pre-assembled vertical drains were carved
of Santa Catarina.
using the spiking of a keyless chuck with 15 cm
The location of the experimental embankments diameter, with a horizontal spacing between
was determined to obtain the lowest possible drains of in a triangular pattern and length
thickness of hydraulic embankment, to simulate in order to cross the full layer of clay.
the most critical situation, looking for the area of
The AE1 embankment featured a length of m,
maximum possible homogeneity both on
plus the edges, with of cross section and
geotechnical and geometric level, to favor the
slopes with an 1,5:1,0 (H:V) inclination.
understanding of the observed data, and finally,
as close as possible to the construction site to The embankment stopped rising in height when it
ensure greater representation and less distance reached the rupture, which occurred at .
to transport the materials which would be used. The performance is presented in Fig. 3.
The obtained location is presented in Fig.2

Fig. 2 – Location of experimental embankment [2]

The embankment AE1 was built with synthetic
reinforcement in the base and the foundation
treated with vertical drains, to examine the
influence of short-term drains on the performance
of reinforced embankments. The main function of
the reinforcements and drains are to improve Fig. 3 – Performance of the embankment AE1 [2]

stability of the embankment body and reduce the

post–building constructive settlements,
respectively. Under a layer of hydraulic
embankment measuring and thick
of soft clay, based on a dense sand stratum.

2
 Carmo Cardoso

The Approach Embankment Bridges Table 1 – Characteristic Parameters of Fine Sand -

Hardening Soil model [1]
The approach embankment bridge is the Parameters Value
structures that connects the bridge with the Geotechnical parameters
ground, with the main objective of reaching the Bulk weight γ [kN/m3] 17,5
required quota and make the connection in safe Initial void ratio e0 0,6
OCR 1
and comfort conditions. The occurrence of -4
Vertical permeability 10
settlements in these infrastructures in the -4
Horizontal permeability 10
procedure phase do not represent by itself risks Constitutive parameters
reaching the Ultimate Limit State, though Effective Friction angle ∅’ [°] 33,8
generates high maintenance costs to ensure the Dilatancy angle [°] 3,8
(*)
Effective Cohesion c’ [kPa] 0
Limit of Use State. 2
Stiffness Modulus [kN/m ] 18 000
2
Stiffness Modulus [kN/m ] 52 000
Selecting the most appropriate constructive
Power for stress-dependency level, 0,65
method is conditioned by several factors such as (*) In the modeling of unsaturated embankment was
the deposit geotechnical features, use of the area adopted c = 1 kPa.

and its surrounding area, construction deadlines

Along with the sand of the draining mattress,
and costs involved. The issue in question has no
applied on the basis of all embankments,
restrictions on the available space and difficulties
modeled with the Hardening Soil material model,
within the neighborhood. Regarding the duration
with the characteristics presented in table 2.
of construction it was agreed on a maximum of
18 months (a year and a half). The purpose of Table 2 – Characteristic Parameters of Sand draining
mattress –Hardening Soil model [1]
this site is to reach a height of above ground Parameters Value
level, a quota and wide, for later Geotechnical parameters
construction of a two-way communication via Bulk weight, γ [kN/m3] 17,4
fixed and their respective edges, with both ramps Initial void ratio, e0 0,5
-4
Vertical permeability, 10
having equal inclination, 2:1 (H:V), as shown in -4
Horizontal permeability, 10
Fig. 4. Constitutive parameters
Effective Friction angle, ∅’[°] 33,8
Dilatancy angle, [°] 3,8
Effective Cohesion, c’ [kPa] 0
2
Stiffness Modulus, [kN/m ] 100 000
Fig. 4 - Section type, in meters Poisson coefficient, ν' 0,30

As in the case of experimental embankment, it is The clay stratum is modeled with the material
considered that the body of the embankment is model of Soft Soil, which parameters are
executed with the same hydraulic fill sand, with presented in table 3. The groundwork layer is
parameters presented in Table 1, and modeled in modeled with Mohr-Coulomb model and material
the software model Plaxis Hardening Soil with the characteristics presented in table 4.
material.

3
 Carmo Cardoso

Table 3 – Characteristic Parameters of Soft Soil - Soft Soil properties due to phased construction. To
model [1]
Parameters Value achieve this the equation of repressions for a
Geotechnical parameters fixed quota of landfill was applied considering the
Bulk weight γ [kN/m3] 13,7 submersion of the layer, by [3], EQ. [A].
OCR 1
-9
Vertical permeability 10 ( ) [A]
-9
Horizontal permeability 10
Constitutive parameters
Having reached a rating of to a
Initial void ratio e0 3,5
Effective Friction angle ∅’ [°] 30 embankment of height .
Dilatancy angle [°] 0
Effective Cohesion c’ [kPa] 5
The secondary settlements calculation based on
Modified compression index λ* 0,1565 the methodology proposed by Martins (2005),
Modified swelling index k* 0,0344 exposed in [2], was applied and reached an
estimation from Eq.[B] a value of .
Table 4 – Characteristic Parameters of Bottom Sand –
Mohr-Coulomb model [1]
Parameters Value ( ) ( ) [B]
Geotechnical parameters
Bulk weight γ 17,5
The conclusion was that it is necessary to build a
Initial void ratio e0 0,5
-4 embankment of height
Vertical permeability 10
Horizontal permeability 10
-4 to compensate the repressions in the
Constitutive parameters completion, .
Effective Friction angle ∅’[°] 33,8
Dilatancy angle [°] 3,8 The overhead is estimated so that the height of
Effective Cohesion c’ 0
the embankment overhead causes a primary
Stiffness Modulus 100 000
discharge equal to or greater than the total
Poisson coefficient, ν' 0,30
discharge previously estimated. It has been
estimated an overload of , using the
Projects
equation of repression by primary densification,
In all projects developed, such as the AE1
EQ. [C], to be removed at the end of the 18
landfill, it was agreed to put a geo-reinforcement
months construction phase.
on the foundation as a backup step.
[ ( ) ( )] [C]
Conventional Embankment
The conventional embankment project is built in
The assessment was to use the same system of
stages and foresees the placement of vertical
drains that were used in embankment AE1.
drains and overhead in order to speed up primary
Mattress sizing draining made according to
repressions and anticipate secondary
Cadergren methodology (1967) exposed in [2],
repressions.
follows with this spacing of drains and with

In the pre-sizing phase, through analytical thick layer of draining able to set seven horizontal

calculation, the primary settlements for the drains spaced by .

simplest situation were forecasted, i.e., it was not

It is estimated the active tractive effort in
considered the improvement of the foundation
strengthening through the methodology proposed
1
smear is the effect of compaction in the soil surrounding the drain, caused by their spiking.
4
 Carmo Cardoso

in 1985, by Rowe and Sodermenn exposed in [2], [5] ,Fig.6. The obtained permeabilities are in the
developed to evaluate the tensile forces following table.
mobilized in strengthening from the value of Table 5 – Permeabilities of Soft Clay
Parameters Value
predictable deformation, , in function of a
Vertical permeability
dimensionless parameter Ω, which relates from a
Horizontal permeability
abacus. Horizontal permeability in the smear
zone
Plane strain horizontal permeability
Through the analysis we projected a strain
for a geo reinforcement presenting a Plain strain horizontal permeability in
the smear zone
tensile stiffness modulus in direction, to 5%
deformation (J5% ), de approximately
[4], we may concluded that the maximum traction
which can be mobilized in the reinforcement is in
an order of magnitude of , much lower than
the .

A predication was made on how long it takes for

the primary densify to occur due to construction
of embankment considering densification with Fig. 6 – Transformation of permeability [5]

purely radial drainage, , due to the effect By this means and after running the model in
of vertical drains. It wasn’t considered the effect Plaxis 2D software, it was possible to get the
1
of smear and changes of the values of the offsets at the top of the embankment which is
coefficients of permeability due to phased presented in graphic form in Figure 6, where it is
construction in order to simplify and to visible a near blistering when the overload
compensate the non-consideration of vertical is removed. It was found that the geo-
drainage component. reinforcement has a maximum axial effort

We decided to make a Plane strain modeling. .

Obtaining the model's figure.

Fig. 5 – Transverse profile of the Conventional

Embankment: 1 - Embankment; 2 - Draining Mattress; 3 -
Hydraulic Embankment; 4 - Clayey Stratum; 5 - Stratus
Bottom Sand

To build the model in Plane strain and to

Fig. 7 – Progress of the settlement on the top of the
represent the smear effect, the change of 2D and embankment and a height of embankment constructed
time-related
3D problem altered the permeability along drains,
by applying the methodology recommended by

5
 Carmo Cardoso

Fig. 7 shows that the construction in stages Table 6 – Characteristic parameters of the materials of
the Stone-Column - Mohr-Coulomb model
produced settlements smaller than estimated
Parameters Value
analytically and that more time is needed to Geotechnical parameters
stabilize due to the densification of clayey Soil unit weight above phreatic level γh
18
[kN/m3]
stratum. Soil unit weight below phreatic level γsat
20
[kN/m3]
-4
Jacketed Stone-Columns Embankment Vertical permeability 1,16x10
-4
Horizontal permeability 1,16x10
The Jacketed Stone-Columns Embankment has
Constitutive parameters
as an advantage regarding conventional Initial void ratio, e0 0,5
embankments: shorter construction time and Effective Friction angle, ∅’[°] 38
added control over the repressions, causing an Dilatancy angle, [°] 12
Effective Cohesion, c’ [kPa] 5
increase in resistance of the clay due to the 3
Stiffness Modulus E’[kN/m2] 25x10
drainage reached and the increased strength of
index n’ 0,30
the foundation.
In the modeling process the unit cell concept was
It was agreed that the columns would measure
used with equivalent diameter ,
length, to be built-in the layer, diameter
using the option of symmetry, Axisymmetry
, forming a square weave of side, in
model.
order to ensure the absence of differential
settlements at the top of the embankment. The
criteria applied was the one proposed by [6]
which presents the EQ. [D].

[D]

Where is the diameter on the top of the column

and is half the highest horizontal distance of
soil.

Imposing height, , with diameter

getting chose to stipulate
, common value in this type of works, for
drainage issues. The granular material used has Fig. 8 – Unit cell and finite element mesh of the Jacketed
Stone-Columns Embankment
the same properties of common sand, modeled
Finding the curve of the settlements at the top of
following the model of material Morh-Coulomb
the embankment on the non-jacketed and
and with the characteristic parameters shown in
jacketed cases, making it clear the contribution of
table 6. All other materials remained
the coating, Fig.9.
identical.Fig.8.

6
 Carmo Cardoso

Table 7 – Characteristic parameters of Concrete

Parameters Value
Bulk weight,, γ [kN/m3] 25,0
6
Stiffness Modulus,
2
[kN/m ] 30x10

Poisson coefficient, ν' 0,00

-4
Vertical permeability, 10
-4
Horizontal permeability, 10

It was observed that the stake had a vertical

Fig. 9 – Progress of the settlement on the top of the displacement of and there are no
embankment and a height of embankment constructed
time-related for 1 – Stone-Column and 2 - Jacketed differential settlements at the top.
Stone-Column

Embankment over piles

In this solution it was considered the

embankment on stands, spiked with
∅ , and with circular capitel with
diameter, forming with a quadrangular
mesh aside , this dimension guarantees
the non-existence of differential settlements at
the top of the embankment, as in the project of
the embankment on granular columns, respects
the criteria proposed by Kempfert in 2004 present
at [2] and ensures that the stresses in the
reinforcement are admissible.

Fig. 10 – Unit cell and finite element mesh of the staked

As in the case of embankment on granular embankment
columns, the modeling of the embankment on
stilts resorts to a unit cell, Fig.10. However, the
axisymmetric modeling does not consider the Economic Analysis
anisotropy of the geo-reinforcement and even
In table 8 the activities and quantities for each
allows checking the stress concentration in the constructive option method is described and at
corners of the capitals. To bypass this limitation, Fig.11 is present the . The quantities shown are
accounted for by embankment and longitudinal
in the sizing phase it was foreseen the
metro as a result, the final prices also refer to
construction of circular capitals without angular €/(underground longitudinal embankment ).
areas.
The unit price values applied to each task are
The concrete stake was designed with a linear representative of the average market values of
elastic behavior, with the properties shown in the the Portuguese market.
following table. All other materials had their
In the proposal for the conventional embankment
properties and models unchanged.
the construction price for the transition paving

7
 Carmo Cardoso

stone was not introduced since this appraisal was The most expensive solution is the proposed
not viable per meter of longitudinal embankment. Embankment on Jacketed Stone-Columns, with a
cost of 5.690,00 € per meter longitudinal
After obtaining the costs for each submitted
embankment, showing no advantage for the
proposal it is reasonable to conclude that the
solution of the Landfill staked. The proposed
more flexible solution - Conventional
Embankment staked is the most rigid and more
Embankment - is in fact the more cost efficient,
reliably.
with a cost of 3.061,00 € per meter of longitudinal
embankment. However it is also the solution that In several projects the incline of slopes did not
requires more time - 18 months - for the change, 1:2 (vertical: horizontal) for being one
stabilization of compressible stratum. initial condition. However, it is more likely that
more rigid solutions may be executed with
The time factor should be analyzed in sets with
steeper slopes, which would translate into less
other factors such as the time of construction of
earth volume that in the end would reflect
other infrastructure or other interventions that the
significantly on the final cost.
area will suffer. The waiting time between stages
of construction could be used, in its entirety, for
the construction of other infrastructure.

Table 8 – Resume of the Economic Analysis

Jacketed
Conventional Embankment
Stone-Columns
Description Embankment on piles
Embankment
Quantity

1. Foundation treatment including raw material price and

all work necessary for its proper application. 392,0 m 120,0 m 52,0 m
2. Installation geo-renforcemt including raw material price
2 2 2
and all work necessary for its proper application. 48,0 m 42,0 m 41,0 m
3. Construction of the embankment, including on the price
3
raw material, quantity of water and its transport, all the 270,0 m
3 3 3
necessary work to compaction and testing. -33,8 m 170,5 m 154,0 m
Total 3.061,00 € 5.690,00 € 4.215,00 €

Fig. 11 – Section type of each solution: A - Conventional Embankment; B - Embankment over Jacketed Stone-Columns;
C - Embankment over piles.

8
 Carmo Cardoso

of the peculiarities overcome in each solution is a

Conclusion
critical factor, in other words if that project can be
Embankments on soft soils can be constructed
extremely careful but the prospect does not allow
successfully if the solution was taken in account
the correct geotechnical characterization, the
the complex behavior of clayey soils.
final result will not be satisfactory.
The solution of the Conventional Embankment,
as the experimental embankment AE1, explores
at most benefit of consolidation accelerated by References
combining the drains and overhead. It is noted [1] Sonney, R. (2013). Numerical analysis of
that the first settlements occur identically, i.e., the 3 test embankments on soft ground:
effect of basal reinforcement and
curve of the settlements at the beginning of the
prefabricated vertical drains. Master
loading exhibit similar trends and values. thesis – COPPE - UFRJ. Rio de Janeiro,
However, it was observed that the progress of Brasil.
experimental embankment AE1 compared to
[2] Magnani, H. O. (2006). Comportamento
conventional embankment modeled showed de aterros reforçados sobre solos moles
lower levels of pore pressure due to the temporal levados à rutura. Doctoral dissertation –
COPPE - UFRJ. Rio de Janeiro, Brasil.
spacing, allowing for their dissipation and
consequently to maintain the stability of the [3] Almeida, M. S., & Marques, M. E. (2010).
foundation. The traction strain measured at the Aterros sobre solos moles: Projeto e
desempenho. Sao Paulo: Oficina de
geo-reforcemt before break was T = 32kN, while
texto
in the model the maximum axial strain measured
was T = 12.2 kN. [4] HUESKER. (s.d.). Stabilenka - Woven
Fabrics for Reinforcement and
With the modeling it was verified that the clay Separation

stratum obtained significant improvements, [5] Indraratna, B., Rujikiatkamjorn, C.,

because of phased construction, resulting in Sathananthan, I., Shahin, M. A., &
smaller settlements and consequently reducing Khabbaz, H. (2005). Analytical and
numerical solutions for soft clay
the permeability coefficient, which causes the
consolidation using geosynthetic vertical
need for more time for the dissipation of pore drains with special reference to
pressures, i.e., more time to stabilize. embankments.

The economic analysis was performed with the [6] Filz, G. (2012). Column-Supported
Embankments: Settlement And Load
intention of compare the cost of each solution Transfer. Geotechinical Engineering
and was used the mean value of the Portuguese State Of The Art And Practice - Keynote
market. However, the values contains prices of Lectures From Geocongrees, (p. 54).

skilled labor higher than those applied in South

America, which means that in the accounting of
the time factor requires special attention.

The importance of a good geological and

geotechnical characterization is the key factor for
the good project performance. The interpretation