Sie sind auf Seite 1von 6

The Prediction and Effect of Tunnelling Induced Movements on the Pile Foundations of a Fly-over Bridge: A

Case-study
Dipl.-Ing. Dr. techn. Friedrich Prinzl
Geoconsult Consulting Engineers, Salzburg, Austria
Dipl.-Ing. Alexandre R.A. Gomes
Geoconsult Consulting Engineers, Salzburg, Austria

ABSTRACT: During the construction of the north section of the Bangkok MRT Chaloem Ratchamongkhon
Line, bored tunnels were driven close to the pile foundations of adjacent buildings and bridges at several
locations along the approximately eleven kilometres of the contract. These conditions required the
development of a practicable methodology for the prediction of tunnelling-induced ground movements and
their impact on the pile foundations. This paper presents the application of this methodology for the
particular conditions encountered at the passage of the bored tunnels alongside a Fly-over Bridge. It also
presents the monitoring programme established for the respective tunnel section and a comparison between
the estimated and the actual deformations for the purpose of assessment validation.
1. INTRODUCTION
The northern section of the Bangkok MRT
Chaloem Ratchamongkhon Line comprises about
9 km of 6.3 m twin bored tunnels, two sections of
cut and cover tunnels and 9 cut and cover stations,
constructed within diaphragm walls using top
down construction techniques.
Along the tunnel alignment, bored tunnels
passed close to the pile foundations of buildings,
channel bridges, and fly-over structures.
Therefore, the development of a methodology to
assess the interaction between tunnels and
adjacent piles, allowing the safe evaluation of
their structural performance and the requirement
of protective measures in advance of tunnel
construction, was an issue of mayor relevance.
In order to present this methodology in general
terms, the paper presents a case study comprising
the assessment of the effect of tunnel construction
on the pile foundations of a Fly-over bridge
located at the tunnel section between the Ratchada
Phisek and Lat Phrao Stations.

after breaking through the station wall, from


where they were launched.
Before arriving at Lat Phrao Station, tunnels
were driven parallel to the ramp of the Lat Phrao
Fly over bridge for part of the section, bending
and crossing the bridge diagonally, as shown in
the Figure 1a.

Figure 1a Tunnel Section along the Lat Phrao


Fly-Over Bridge
N

FLY OVER
BRIDGE

2. RATACHADA PISEK - LAT PHRAO


SECTION
Bored tunnels were driven with the use of two
Earth Pressure Balance (EPB) shield machines.
The tunnel lining was constituted of precast
concrete segments (segmental lining) with an
outer diameter of 6.3 m and a thickness of 0.30m.
Bored tunnels were driven from Ratchada Pisek
towards Lat Phrao Station with an interval of one
month. Along this section the tunnels encountered
the foundation piles of a Channel Bridge shortly

BORED
TUNNEL
AXES

Figure 1b View of the Ratchada Pisek Road and


the Lat Phrao Fly-Over Bridge

The foundations of the Channel Bridge


comprised of five rows of piers, consisting of two
series of sixteen driven piles of 0.35x0.35m crosssection. Before the TBM passage, the bridge was
underpinned by the installation of new bored piles
of 1.2m diameter. The existing piles within the
tunnel cross-section were pulled out from ground
surface or had to be cut manually and removed
from the shield chamber, when the first procedure
failed. Figure 2 shows the general arrangement at
the channel bridge section.

3. GROUND CONDITIONS
Soil conditions along the section correspond to
the typical Bangkok subsoil, which is constituted
of a first layer of man made ground of about 2.5m,
followed by a soft clay layer 11 to 15m thick
underlain by a stiff clay layer with thickness
ranging from 6 to 14m. Below these clay layers is
the topmost Bangkok aquifer, which is constituted
of dense to very dense sands interlayered by thin
layers and lenses of stiff to hard clays.
Along the Fly-over section, bored tunnels were
driven mostly in the stiff clay layer. However,
along a short section, tunnels were also partially
embedded in the underlying sand layer.
A typical arrangement at the fly-over section
including the soil profile is shown in Figure 3.
4. ASSESSMENT OF GROUND MOVEMENTS

Figure 2 Tunnel passage through the Channel


Bridge
The Fly-over bridge foundation consisted of
single reinforced concrete piles located along the
bridge approach ramp and groups of four to six
piles for the bridge piers. All deep foundation
consisted of 0.80m diameter bored piles.
Due to the bend of the alignment, piles
encountered during tunnel drive were cut-off and
removed manually from the chamber of the shield
machine. Bored tunnels passed as close as 1.6m
from the rim of other remaining pile shafts.

Figure 3 Tunnel passage along Fly-over Bridge

The tail void between the segmental lining and


the excavation diameter of the shield machine was
approximately 65mm. Although this gap would be
continuously back-grouted during the shield
advance, some soil displacements would
inevitably occur towards the cavity (ground loss),
inducing horizontal and vertical movements. Also,
additional ground losses would take place as a
result of face losses caused by insufficient
pressure at the shield chamber.
In view of the heterogeneous ground
stratigraphy, it was not possible to use empirical
or analytical methods, such as that proposed by
Loganathan and Poulos (1999), to assess
tunnelling induced ground movements at the
depth. Therefore, in order to estimate the
magnitude and form of these displacements, a
series of analyses was carried out at typical pile
locations, i.e. between 1m to 12m distant from the
tunnel lining, with the use of a two-dimensional
explicit finite difference method (FLAC code,
Cundall et al 1993). With this numerical method, a
complete ground/structure interaction could be
modelled, including both stress and strain
distribution in the ground and deformation and
section forces of the lining.
The ground behaviour was modelled on the
basis of a perfect elasto-plastic constitutive model
and volume losses due to tunnelling were
simulated by allowing stress relaxation and
displacements at the tunnel boundaries. The stress
relaxation used in the analysis was defined
interactively so that it would correspond to the
upper bound ground loss expected for this tunnel
section, which was in the order of 2%. As to
consider the non linear relationship between
ground stiffness and strain, a higher Youngs

Modulus was used outside a zone of about half of


the Tunnel diameter (Prinzl and Gomes 1998).
5. EFFECT ON THE PILE FOUNDATIONS
Regarding the new pile foundations of the
Channel Bridge, additional section forces caused
by tunnelling induced soil displacements were
considered in the design, so that sufficient reserve
was allowed to accommodate these effects.
On the other hand, since the pile foundations of
the Fly-over Bridge were already installed without
considering these additional section forces, a
detailed evaluation had to be carried out.
Details of the Lat Phrao Fly-over Bridge
foundations were available from the respective
design drawings. Piles were located adjacent to
and with their tips below the tunnels. The concrete
cube strength was 30MPa. Reinforcement rates
varied both among the different piles as along its
depth.
Due to the complexity involved in the
simulation of a large number of different analysis
cases with the FD method, an indirect approach
was chosen to allow an easier consideration of the
different pile locations and dimensions.
Hence, the interaction between tunnelling
induced movements and pile foundations were
assessed with a combined analysis method, which
considered separately the effect of horizontal and
vertical ground movements, as explained in the
following sections 5.1 and 5.2.
5.1

Horizontal Movements

Horizontal ground movements due to tunnelling


may cause bending moments on the piles. This
effect was analysed on the basis of the
displacements determined by the FD method and
the use of the program ALP, for laterally loaded
pile analysis (OASYS ALP, 1997). In this
analysis, the pile was modelled as a series of
elastic beam elements and the soil as a series of
non-interactive, non-linear springs. The soil-load
deflection was modelled by using a pre-defined
horizontal soil deformation (P-Y curve), which
was taken from the ground movements assessed
with the FD method.
Generally, the effect of tunnelling induced
horizontal movements would be of larger
magnitude for longer piles and piles closer to the
tunnel. Therefore, the effect of ground movements
on the piles was analysed more in detail for
critical cases, where the distance between the pile
axis and the two tunnels was minimum and the
piles deeper.
Figure 4 shows the results of a analysis with the
ALP program for the case of ground movements

due to both bored tunnels assessed with the FD


method for the analysis of a 39.0m deep pile,
located 1.6m from the outer lining of the nearest
tunnel.

SOIL
DISPLACEMENTS

BENDING MOMENTS

PILE
DISPLACEMENTS

Figure 4 Bending Moments caused by Ground


Movements as calculated by the ALP software
The total section forces in the pile would result
from the superposition of tunnelling induced
forces with section forces due to loads transferred
from the upper structure. Since no information
was available on the design bridge loads, an
estimation was performed to establish the possible
existing section forces acting on the piles.
The estimation has shown that existing
horizontal loads in the pile caps were minor,
resulting in small bending moments at greater pile
depth. Therefore, critical conditions due to the
superposition of existing with tunnelling induced
bending moments were considered unlikely.
As no existing bending moments were used in
the assessment, it was considered reasonable to
exclude also axial forces due to traffic loads,
which would introduce a beneficial effect. Thus,
in view of the uncertainties involved in the
assessment, it was generally carried out
conservatively by assuming that piles were under
pure flexure. Bending moments obtained from the
analysis were then compared with the allowable
moments for the minimum reinforcement rates
provided at different pile depths.
Results of this analysis showed that moments
originated by the horizontal displacements could
be reasonably covered by the as-built
reinforcement.

Working Load

Settlements (mm)

Settlements

Settlements induced by tunnelling can create


negative skin friction along the pile, which would
impose additional load and settlements that could
cause damage of the upper structure.
Due to the ongoing land subsidence in
Bangkok, piles are generally designed to
withstand this additional load. However, whilst
the land subsidence is a long-term process and is
equally distributed, the negative friction forces
originated by ground settlement from tunnelling
occurs in a relatively short period and their
magnitude depends on the distance between the
pile and the tunnel.
For the deep piles of the Fly-over Bridge, the
negative skin friction (that would develop mainly
in the soft clay layer) was found to be
considerably smaller than the frictional resistance
(skin friction) in the lower portions of the pile.
When the negative friction was imposed, axial
thrust in the pile shaft would increase and the pile
would settle activating the positive skin friction
until a new equilibrium was reached.
In this way, maximum settlement would be
limited to the amount that is considered necessary
to activate fully the skin friction resistance.
According to pile testing results, this settlement is
usually in the order of 1% of the pile diameter
(e.g. 8mm for a 0.8m diameter pile).
Such an assessment of pile settlements is shown
for a pile located at the approach ramp of the Flyover, where conditions are more critical due to the
higher working loads. For the assessment, ground
settlements at a distance of 1.6m from the tunnel
were used and a maximum pile settlement of
10mm was assumed (see Figure 5). From the loads
assessed for the bridge, a maximum vertical load
(working load) was back calculated.
According to this assumption, the following
loads would be acting on the pile:

Zone1: Negative skin friction is limited to a


depth of 15.0m (vertical movement of 10mm)
and is fully activated above the depth of 13.0m
(vertical movement of 18mm)
Zone 2: Between 15m and 18m depth, friction
forces are partially activated (pile settles more
than the surrounding ground).
Zone 3: Friction forces are fully activated from
a depth below 18.0m (correspondent to a
vertical soil movement of 2mm).

As a conservative assumption, the negative skin


friction was assumed to a depth of 15m depth,
whereas positive skin friction was applied only
below 18m depth.

-26 -24 -22 -20 -18 -16 -14 -12 -10 -8

-6

-4

-2 0
-5
-10
-15

4
Zone 1

Zone 2

-20
- Zone 1: Negative Skin Friction

Depth (m)

5.2

-25

- Zone 2: Positive Skin Friction not fully


mobilised (neglected)
- Zone 3: Positive Skin Friction fully mobilised

Zone 3

-30
-35
-40
-45

Assumed Pile Settlement = 10 mm


-50

Figure 5 Ground Movements and Sections


forces at the Pile
In this analysis, the comparison between
negative and positive forces has to satisfy the
following conditions:
Qpos
Qneg + WL

, where:

Qpospositive skin friction forces


safety factor (assumed as 1.5)
Qnegnegative skin friction forces
WLworking load
Results showed that since the reduced positive
skin friction forces are larger than the sum of the
negative skin friction forces and the working load,
pile settlements would be limited to 10mm.
According to the structural and visual evaluation
of the bridge, this magnitude of differential
deformation was not considered to be critical for
the super structures, which consisted mostly of
piers and single span beams in good structural
conditions. Furthermore, this deformation could
also be easily corrected by adjustment of the
bridge bearings, if so required.
6.
OBSERVATION DURING
CONSTRUCTION
6.1

Monitoring Program

Based on the results of the assessment, it was


decided that auxiliary and/or advanced protective
measures were not fundamentally required for the
Fly-over Bridge, providing that ground losses
could be properly controlled during the passage of
the tunnels.
Indeed, a damage or local overstressing of a pile
could not be totally discarded, as the interaction
would be controlled by a series of factors that
could adversely impact the piles, such as:

Closer proximity to the piles because of


deviations during tunnel drive and actual
pile position.
Changes in the shield drive performance
(excessive ground losses)
Unidentified bridge structural damages
and/or deviations in the material properties
or as-built characteristics of the piles.

Therefore, in order to allow a controlled tunnel


drive along the bridge section, an observational
approach was applied to allow a close control of
the driving performance.
This observation approach included a
comprehensive monitoring of the horizontal and
vertical ground deformations ahead of tunnel
driving through critical sections and the backanalysis of the results of the geotechnical
monitoring.
The monitoring programme required the
installation of the following instruments:

Surface and sub-surface settlement


markers installed in pairs along the two
tunnel axes, spaced between 20 to 50m in
accordance of the specific requirements
along the section.

Additional surface settlement markers


installed above the axes with 10 m spacing
ahead and along the critical sections of the
channel and the fly-over bridges.

Combined inclinometers-extensometers for


the monitoring of horizontal displacements
and settlements at the depth, in advance of
and close to critical pile foundations.

Vibrating wire piezometers to assess porepressure changes due to tunnelling drive.

Convergency bolts to control the tunnel


lining deformation.

Settlement markers at all pile caps,


columns and the upper structure of the
bridge for monitoring of its structural
response.

allow the early implementation of modifications in


the construction performance (e.g. change of
pressure in shield chamber and sequence of tail
void grouting), and the early provision of
protective measures.
6.2

Monitoring Results

Settlements back calculated from the


monitoring results have shown that the drive
performance improved with tunnel progress.
Actually, back calculated ground losses reached
up to 2.5% at the early section of the first tunnel
drive, reducing gradually to about 1.5% before
reaching the Channel Bridge.
This improvement was partially related to the
lowering of the tunnel alignment, which changed
from a mixed face condition, where the tunnel
crown was in the soft clay, to be fully embedded
in the stiff clay layer or with its invert at the sand
layer. Additionally, it could also be clearly related
to the learning curve of the shield drive and
experience of the crew with the actual ground
conditions.
As expected, the second tunnel drive showed a
far better performance, with back calculated
ground losses showing magnitudes of less than
1.0%.
The passage through the Channel Bridge
showed no major surprises. Actual displacements
were less than considered by the design and pile
stresses were always within tolerable limits,
showing no sign of damage.
Figure 6 shows a comparison between the
estimated horizontal movements from both tunnels
and the readings of inclinometers 26-IE-001 and
002 installed close to the new bored pile
foundations and spaced approximately 1 to 2m
from the tunnel.
Horizontal Displacements (mm)

0
-30

-25

-20

-15

-10

-5

10

-5

-10

-15

Depth (m)

Additionally, visual inspection of the bridge


structure for the appearance of cracks during
tunnel drive was required.
Three trigger levels were defined for the
monitoring readings (alert, alarm and action
level), to provide a continuous and systematic
comparison between the actual ground movements
and the design assumptions, which considered a
maximum value of 2% ground loss.
The attainment of a specific trigger level would
trigger certain pre-established actions by the
responsible parties (Monitoring Team, Contractor,
Designer), providing the landmarks for a
controlled drive along the more critical sections.
The identification of critical movements should

-20
Distance to closest tunnel
FD analysis +1.0m

-25

FD analysis +2.0m
Inclinometer 26-IE-001

-30

Inclinometer 26-IE-002

-35

Figure 6 Estimated horizontal movements due


to two tunnels (FD Analysis - 2% ground loss) and
readings of inclinometers 26-IE-001 and 002.

With the experience gained along the initial


sections, tunnel drive continued to improve its
performance, as shown by the back calculated
ground losses, which reduced to only about 0.5%
for the section along the Fly-over bridge.
Instruments installed along this section,
recorded readings with maximum horizontal
displacements (inclinometers) of about 5mm,
against the estimated 12.5mm.
Whereas the form of the monitored horizontal
displacements confirmed the results of the FD
analysis, the pattern of monitored settlements did
not correspond to the expected pattern of
settlements in the depth. This was probably caused
either by a malfunction of the extensometers,
which may not have been properly installed or
were just due to the very small magnitude of the
movements and insufficient accuracy of
monitoring.
At any extend, surface settlements could be
confirmed by the installed surface and sub-surface
settlement markers, which registered peak
settlements at the surface smaller than 15mm. This
value was much lower than the estimation, which
predicted maximum settlements of about 40mm at
the alignment axis (between the two tunnels).
With base on the observed movements, it could
be concluded that actual volume loss, which was
in the order of 0.5% ground loss, was of a
magnitude considerable inferior than that assumed
for the pile impact assessment.
Regarding the bridge response, instruments
installed at the pile caps and barrettes of the Flyover showed settlement/heave in the order of
about 5mm which is about half of the predicted
value.

The Figure above shows a comparison between


horizontal displacements obtained from a FD
analysis (two tunnels, 0.5% ground loss) and the
readings of inclinometers 26-IE-005 and 006,
which were installed close to the pile foundations
and spaced approximately 1 to 2m from the
tunnel.
7. CONCLUSIONS
A methodology for the prediction of tunnellinginduced ground movements and their impact on
pile foundations was outlined by showing the
assessment and results of an actual tunnel drive
alongside a Fly-over Bridge.
Based on this assessment, recommendations
could be made regarding the risks involved in the
construction and the requirement of protective
measures. Additionally, it also provided the basis
for the establishment of a specific monitoring
programme for the respective tunnel section,
including the definition of acceptable limits of
deformation and structural response.
Results of the monitoring have shown that by
means of a controlled and careful drive, ground
losses could be maintained below the design
assumptions, causing no harmful effect to the pile
foundations of the adjacent structure, as observed
on site.
8. ACKNOWLEDGEMENTS
The authors would like to express their gratitude
to the Metropolitan Rapid Transit Authority
(MRTA), the ION Joint-Venture and Ove Arup
and Partners for the possibility of publishing this
paper.

Horizontal Displacements (mm)


0

-5

-4

-3

-2

-1

-5

-10

Depth (m)

-15

-20

-25
Distance to closest tunnel
FD analysis +1.0m
FD analysis +2.0m

-30

Inclinometer 26-IE-005
Inclinometer 26-IE-006

-35

Figure 7 Horizontal displacements due to the


two tunnels (FD analysis - 0.5% ground loss) and
readings of inclinometers 26-IE-005 and 006.

9. REFERENCES
Cundall P.A., Coetze M.J., Hart R.D., Varona P.M. (1993)
Flac Users Manual. Itasca Consulting Group. U.S.A.
Bangkok Metropolitan Administration 10 Overpass
bridges Site 9 Ratchadapisek Lat Phrao Intersection
Drawings H- and S- Series. March 1993.
Loganathan N. and Poulos H.G. (1999), Tunnelling Induced
Ground Deformations and their effect on Adjacent Piles,
The Race for Space - 10th Australian Tunnelling
Conference, March 1999, Melbourne, Australia.
OASYS ALP User Manual Version 3.4 Laterally Loaded
Pile Analysis. 1997.
Prinzl F. and Gomes A.R.A., (1998). The Requirement of
Protective Measures for Buildings affected by Ground
Movements due to Shield Tunnelling in Soft Ground.
ITA World Tunnel Congress 98 on Tunnels and
Metropolises, April 1998, Sao Paulo, Brazil.

Das könnte Ihnen auch gefallen