Professional Documents
Culture Documents
Ground Improvement Techniques For Railway Embankments - 2012
Ground Improvement Techniques For Railway Embankments - 2012
Civil Engineers
Ground Improvement 162
February 2009 Issue GI1
Pages 3–14
doi: 10.1680/grim.2009.162.1.3 Arul Arulrajah Affendi Abdullah Myint Win Bo Abdelmalek Bouazza
Senior Lecturer, Civil Technical Consultant, Director (Geo-Services), Associate Professor,
Paper 800013 Engineering, Faculty of Ranhill Consulting, Kuala DST Consulting Engineers Inc., Department of Civil
Received 01/11/2006 Engineering and Industrial Lumpur Malaysia Ontario, Canada Engineering, Monash
Accepted 12/03/2008 Sciences, Swinburne University, Melbourne,
University of Technology, Australia
Keywords: embankments/railway Melbourne, Australia
systems/rehabilitation, reclamation
& renovation
A high-speed railway project for trains of speeds of up to P peak angle of shear stress
160 km/h is currently being constructed between Rawang Pall coll allowable load on the column
and Bidor (110 km long) in Peninsular Malaysia. The s spacing of the columns in square grid pattern
ground improvement methods adopted in the project T shear stress
are vibro-replacement with stone columns, dry deep soil Th time factor for consolidation by horizontal drainage
mixing (cement columns), geogrid-reinforced piled t time
embankments with individual pile caps and removal/ U degree of consolidation
replacement works. This paper provides a detailed Us degree of settlement
insight into the design and implementation of vibro- Up average degree of pore pressure dissipation
replacement and the deep soil mixing treatment v volume of the element
methods used in the project. The use of plate bearing imp imposed stress due to dead load and live load on top
tests and field instrumentation to monitor the of the ground surface
performance of the stone columns and soil mixing ªc unit weight of column
ground treatment methods is also discussed. This paper ªf unit weight of fill
also provides a brief overview of other treatment allowable strain (ultimate)
methods implemented in this high-speed railway project all col allowable creep stress
such as a pile embankment with geogrids and removal/ fcol failure stress
replacement works. creep col creep stress of column
h horizontal stress calculated at the top of soft layer
NOTATION v9ert vertical stress
Acol /A area ratio d col undrained shear stress of the column
Acol area of column u col undrained shear stress of the column
a area replacement ratio calculated as a ¼ (Acol /s 2 ) for 9 friction angle of the composite system
square grid pattern of spacing, s c friction angle of column
C cohesion of the composite system d col drained angle of friction of the column
Cu undrained shear strength of the soil d comp drained angle of friction of composite soil
Cd comp drained cohesion of composite soil u comp undrained angle of friction of composite soil
Cu comp undrained shear strength of composite soil d soil drained angle of friction of the in situ soil
ccreep col creep stress of the column s friction angle of the soil layer
cd col drained cohesion of columns u col undrained angle of friction of columns
cd soil drained cohesion of in situ soil u soil undrained angle of friction of in situ oil
ch coefficient of consolidation for horizontal flow
cu col undrained cohesion of the column 1. INTRODUCTION
cu soil undrained cohesion of the in situ soil The electrified high-speed railway project runs between
Dc constrained moduli of columns Rawang in the state of Selangor and Bidor in the state of Perak
Ds constrained moduli of soil in Peninsular Malaysia over a total length of 110 km. Figure 1
d diameter of column indicates the location of the project site in Peninsular Malaysia.
de diameter of the equivalent soil cylinder The geotechnical design of the project includes ground
Ecol Young’s modulus of the columns improvement of the existing foundation to sustain the imposed
Ecomp Young’s modulus of composite soil dead and traffic loads for train speeds of up to 160 km/h. The
Esoil Young’s modulus of the in situ soil client’s design requirements are a maximum post-construction
m9 proportional load on the column settlement of 25 mm in six months and a differential
mc constant settlement of 10 mm over a chord spanning 10 m. In addition,
mE constant the degree of consolidation to be achieved is not to be lower
n2 final improvement factor than 85–90%. The required minimum long-term factor of
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 3
ST
KUALA
S
STONE COLUMNS
OF
RA
LUMPUR
IT
Vibro-replacement with stone
MA
S SELANGOR
LA
OF columns is a subsoil
RAWANG (km 355·5)
CC
MA
improvement method in
A
LA
PENINSULAR SINGAPORE which large-sized columns of
CC
KUALA LUMPUR
MALAYSIA
A
coarse backfill material are
installed in the soil by means
of special depth vibrators.
The stone columns and the
intervening soils form an
Figure 1. Location of project site in Peninsular Malaysia integrated foundation support
system having low
compressibility and improved
safety for slope stability was 1.5. Due to the stringent load-bearing capacity. Vibro-replacement with stone columns
settlement restrictions and the fast-track nature of the project, allows for the treatment of a wide range of soils, from soft
an array of ground improvement techniques had to be clays to loose sands, by forming reinforcing elements of low
implemented in locations with soft soils or loose sands on compressibility and high shear strength. In addition to
which proposed high embankments were identified. Ground improving strength and deformation properties, stone columns
improvement was thus required to ensure adequate densify in situ soil, rapidly drain the generated excess pore
performance of the embankments in terms of settlement and water pressures, accelerate consolidation and minimise post-
slope stability as well as completion of the project within the construction settlement. Normally the columns fully penetrate
required project duration. the weak layer with the result that the stone column and
natural soil combination develops greatly enhanced bearing
This paper provides a detailed insight into the vibro capacity and reduced compressibility characteristics. The
replacement with stone columns and dry deep soil mixing method is an ideal solution for use in embankments as it
treatment methods applied in the project. Vibro replacement negates the effect of a ‘hard point’. The dry or wet method of
with stone columns is a subsoil improvement method in which installation can be used depending on the proximity to the
large-sized columns of coarse backfill material are installed in existing railway track and water sources. The size of the
the soil by means of special depth vibrators. Dry deep soil vibrator is around 40 cm and penetration of the vibrator into
mixing technology is a development of the lime–cement the ground with water jetting will result in a hole of diameter
column method. This paper also briefly discusses piled 50–60 cm being created. An annular space is created between
embankments with geogrids and removal/replacement, which the vibrator and the hole through which the stone is fed to the
were also treatment methods adopted in this project. compaction point. The up and down motion of the vibrator is
used to laterally displace the stone into the ground and at the
The railway embankments in the project have heights ranging same time compact the stone column. This will result in the
from 1 to 12 m. The top of the embankment has a minimum creation of the required diameter of column. Figure 2 presents
width of 14.9 m for embankments less than 10 m in height and
a width of 24.9 m for embankments greater than 10 m in
height. The side slopes of the embankments have gradients of
1V:2H. Berms of 3 m width are provided on either side of A
embankments which were greater than 5 m in height. The soils
encountered on the project site are highly variable mixtures of
very soft silts and clays, as well as loose sands to depths of up
to 30 m. Two approaches were needed for the treatment process
due to construction constraints: (a) treatment of the full width
of the embankment was required in locations where the new
alignments needed the construction of two new tracks; (b)
treatment at locations where a new track was to be first
Figure 2. Schematic illustrating the stone column installation
constructed while the existing live track was to be later process (courtesy of Keller)
rehabilitated. Treatment in the first stage would be for the
4 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.
3 Th ¼ 0:044
2.1.5. Design details. Based on the analyses of the stone
column areas, the following design parameters and design
(Balaam and Booker chart for U ¼ 90%, de /d ¼ 3), where t is spacing were adopted
time; Th is the time factor for consolidation by horizontal (a) diameter of column, d ¼ 0.8 to 1.0 m
drainage; de is the diameter of the equivalent soil cylinder; and (b) unit weight of column, ªc ¼ 22 kN/m3
ch is the coefficient of consolidation for horizontal flow. (c) friction angle of column, c ¼ 408
(d) constrained moduli of columns, Dc ¼ 120 MPa
2.1.3. Strength properties of improved ground. Stone columns (e) constrained moduli of soil, Ds ¼ 100 3 Cu ¼ 500 3 SPT
deform until any overload has been transferred to the (f) unit weight of fill, ªf ¼ 20 kN/m3
neighbouring soil. The stone columns receive an increased (g) traffic load ¼ 30 kN/m3 .
0
For soft soils conditions encountered in the Rawang to Bidor
Barron’s solution
Biot theory stretch, stone column spacings were generally in the range 1.8–
0·2
E1/E2 ⫽ 40 10 1 2.3 m for embankment heights of 5–12 m.
0·4 de
Us
d
Predicted total settlements were of the order of 0.3–0.5 m.
Rigid
Up
0·6 Factors of safety for slope stability were greater than 1.5. Time
E1 E2 Raft
v1 v2 required for 90% degree of consolidation in the predominantly
kh sandy silts was less than two months. The treatment area ratio
0·8
Smooth Rigid
for the stone columns varied from 13 to 20%, depending on
v1 ⫽ 0·3, v2 ⫽ 0·3
1·0 the design spacings for the stone columns.
0·0001 0·001 0·01 0·1 1·0
c rt
Th ⫽ 2.2. Stone column installation
d 2e
Arulrajah et al.,5 have described the soil conditions and soil
parameters relevant for stone column design in the project site.
Figure 3. Time rate of settlement of stone columns for
de /d ¼ 3 (Balaam and Booker4 ) The results of site investigations revealed the presence of a
wide range of soils along the track, ranging from very soft silty
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 5
Stone columns were used to treat soils over about 14 km length 2.3. Plate load testing on stone columns
of the railway line. Approximately 1 100 000 linear metres of After completion of stone column installation, plate load tests
0.8–1.0 m diameter stone columns were installed on the project were carried out on single-columns, or groups of four columns
site to depths of 6–30 m. for acceptance purposes. The load was applied on the stone
column and the soil surrounding the column. For the first cycle,
Figure 5 shows the schematic diagram of stone column the allowable design load was applied and maintained for a 24 h
treatment works as carried out in the project at locations of duration. In the second cycle, a maximum load of 150% of the
Tip resistance Qc: MPa Friction Fs: kPa Friction ratio: % Pore pressure Pw: kPa
0 1 2 3 4 5 0 25 50 75 100 0 1 2 3 4 5
0 0 200 400 600 800 1000
0 0 0
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
6 6 6 6
7 7 7 7
8 8 8 8
9
Depth: m
9 9
Depth: m
9
Depth: m
Depth: m
10 10 10 10
11 11 11 11
12 12 12 12
13 13 13 13
14 14 14 14
15 15 15 15
16 16 16 16
17 17 17 17
18 18 18 18
19 19 19 19
20 20 20 20
Figure 4. Plot of typical pre-treatment CPT result at stone column location (chainage 352130)
Embankment S
CL
4950 Min. 5000 4950 1000 mm∅
New track Rehab. track
CL CL 300 mm sub-ballast
1·5 1·5
2 1 1 2
1 1 3000
5·0% 5·0%
Formation
1 m compacted 1 1
1 5000 mm 2
2 sand blanket 2
max.
1000 mm∅
stone column Soft/loose
L material
Embedded in
SPT N ⭓ 10
S
Dense layer
W
Figure 5. Schematic of stone column treatment scheme for new alignment comprising two new tracks (dimensions in mm)
6 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.
Embankment
CL S
4950 Min. 5000 4950
Suitable fill material
compaction in layers
New track Rehab. track 1000 mm∅
CL CL
1 m compacted 300 mm sub-ballast
sand blanket
1·5 1·5
1 1
2
1 2
5·0% 5·0% 1
Existing Fill
Existing
ground
Formation ground
Soft/loose
material L 1000 mm∅
Embedded in
stone column
SPT N ⭓ 10
Dense layer S
Figure 6. Schematic of stone column treatment scheme for partial width treatment next to the existing railway track
(dimensions in mm)
design load was applied. The acceptance requirement of the load surface settlement plates and settlement markers. The surface
test was that the settlement should not exceed 50 mm under the settlement gauges on site consistently indicated settlements
allowable design load and not exceed 80 mm under 150% of the occurring for each additional lift and minimal post-
allowable design load. The size of plate used for the load test was construction settlements. Figure 9 presents typical results from
1.5 m by 1.5 m for a single column and 3 m by 3 m for a group of a settlement plate installed on the project site. Figure 10
four columns. Figure 7 presents the schematic diagram showing presents a typical Asaoka plot for the said settlement plate,
the plate load test set-up for a single column load test. The results which indicates that the degree of consolidation of the
of a typical single column plate load test carried out in the project improved ground at the location had achieved 94%. The long-
are presented in Figure 8. term performance of the stone columns were predicted by
means of the Asaoka method of back-analyses based on the
2.4. Field instrumentation of stone columns field settlement results. Settlement markers were also placed to
Extensive field instrumentation was carried out in the stone monitor the long-term performance of the stone columns after
column treatment areas after the installation of the stone the railway tracks were placed. The long-term monitoring
columns. The majority of the field instrumentation comprised results indicated that the stone columns had performed very
Crane
Crane tracks
Firm fill
Dial gauge Jack
Reference
Top of sand platform beam
1 m nominal diameter
stone column
Figure 7. Schematic diagram showing single column plate load test set-up
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 7
2·0 2·11 mm
2·5
Settlement (i): m
3·0 240
3·5
4·0
4·5 4·2 mm 200
5·0
0 40 80 120 160 200 240 280 320
20 60 100 140 180 220 260 300
Load: kN
160
Degree of consolidation ⫽ St /Sult
Figure 8. Load–settlement curves for a single column plate ⫽ 265·3/282
load test (chainage 352066) ⫽ 94%
120
120 160 200 240 280 320
Settlement (i ⫺ 1): m
6·0
5·5
5·0 Figure 10. Asaoka plot and determination of degree of
Embankment height: m
8 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 9
3.2. Dry deep soil mixing installation 3.4. Field instrumentation of dry deep soil mixing
The results of soil investigation at the treatment area revealed During the construction of the embankment over the treated
the presence of a wide range of soils along the track, ranging ground, settlements and lateral movements of the embankment
from very soft silty clay or clayey silt to loose silty clayey were monitored using rod settlement gauges and inclinometers.
sand. 11 Figure 11 shows a typical CPT plot at the DSM Typical results from the rod settlement gauges are shown in
treatment location. Fig. 15. The settlement gauges showed virtually no settlement
(, 10 mm) for an embankment of height ranging from 1 to
Deep soil mixing treatment was used to treat soft soils over an 1.5 m. Typical results of lateral displacement (perpendicular to
800 m length of the railway line. Over 50 000 linear metres of the alignment of the embankment) are shown in Figure 16. The
0.6 m diameter columns were installed at the site to depths of inclinometers showed lateral movements to be within 15 mm in
6–14 m. 11 The embankment heights in the DSM treatment the direction perpendicular to the embankment alignment. The
areas varied between 1.5–3 m. Column spacing generally inclinometer measurements presented in the figure were
ranges between 1.0–1.5 m. Typically the spacing of the column monitored for seven months. Maximum displacement was
grids (square/rectangle) varies between 1.0–1.3 m centre to observed at the ground level. 11 Details of the field
centre under the location of the proposed rail tracks and 1.4– instrumentation of the dry deep soil mixing works at the
1.5 m centre to centre in the remaining area underneath the project site have been described by Raju. 11
4
Depth: m
10
12
Maximum depth ⫽ 11·00 m Depth increment ⫽ 0·05 m
Figure 11. Plot of typical pre-treatment CPT result at deep soil mixing location (chainage 341690)
10 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.
Working
Existing Embankment
platform
ground
0·6 m dia.
cement
columns
Treatment ⫽ 22·90 m
Figure 12. Schematic of dry deep soil mixing treatment scheme (Raju 11 )
10
0
Grillage beams
⫺10
⫺20
RSG-3 RSG-6
Sand blanket
Cement
Figure 15. Time–settlement plot showing the typical results of
columns rod settlement gauges installed in a deep soil mixing
treatment area
Load: t 4
0 20 40 60 80 100 120
0 8
Depth: m
12
⫺2
Settlement: mm
16
⫺4
20
⫺6
24
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 11
CL Embankment
Proposed formation level
4950 Min. 5000 4950
New track Rehab. track
300 mm sub-ballast
300 mm crusher run 1·5 CL CL 1·5
1 1
5·0% 3000
5·0%
Fill Formation
150 mm sand 1 2 Layers of high strength 1 1
5000
2 2 2 Detail ‘A’
geogrid, Kg 400/200
Direction of track
300 mm Geogrid, kg 400/200
crusher run Top and bottom layer 200 kN/m
1000 overlapping along the track
150 mm sand
Varies
250 mm ⫻ 250 mm
Geogrid, kg 400/200 precast Top and bottom layer 400 kN/m
RC piles Grade 45 perpendicular to the track
12 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.
H
750 mm
Crusher run blanket Geogrids
1 (wrapped with geogrid) Continuous slab
3
L2 ⭓ 6 m
L2 ⭓ 6 m
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
400 mm
1m
1m
S ⫽ 1·5 m RC piles
Piled transition
Civil works Bridge works with individual
pile cap Integrated pile embankment
Articulated slab
Abutment
RC piles with
individual pile caps
750 mm
5 ⫻ pile spacing ⫽ 7·5 m
Figure 18. Schematic diagram showing bridge approach transition using geogrid-reinforced piled embankments
Embankment
CL
Figure 19. Schematic diagram showing typical cross-section of removal/replacement works (dimensions in mm)
have been discussed. The results from numerous load tests and embankments have been discussed. Results from numerous
settlement plates indicate that the stringent performance load tests, settlement plates and inclinometers indicate that the
requirements of the new railway project were met. stringent performance requirements of the new railway project
were met.
Ground improvement by means of dry deep soil mixing allows
for the treatment of a wide range of soils, ranging from soft Piled embankments were designed for use for the bridge
clays to loose sands by forming stronger reinforcing elements approach transitions and allow for the embankments to be
of low compressibility and high shear strength. In this paper, constructed rapidly without any slowdown in the construction
the design methodology, installation methodology, load testing rate or sequence. Piled embankments also eliminate the effect
and field instrumentation for dry deep soil mixing for railway of settlement and stability problems.
Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al. 13
14 Ground Improvement 162 Issue GI1 Ground improvement techniques for railway embankments Arulrajah et al.