th

The 17 Southeast Asian Geotechnical Conference

Taipei, Taiwan, May 10~13, 2010

Instrumented Trial Embankment on Soft Ground at Tokai, State of Kedah,

Malaysia
Yean-Chin TAN1, Lean-Hock OOI2, Andrew YEOW3, Peir-Tien, LEE4 and Kuan-Seng, KOO5
1
Director, G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
2
Head of Geotechnical Department, Gamuda Engineering Sdn Bhd, Kuala Lumpur, Malaysia
3
Senior Manager, Gamuda Engineering Sdn Bhd, Kuala Lumpur, Malaysia
4
Associate, G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
5
Senior Engineer, G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
E-mail: 1 [email protected], 2 [email protected], 3 [email protected],
4
[email protected], 5 [email protected]

ABSTRACT: The geometrical tolerance of railway tracks is generally very stringent, especially at high operating speeds. Hence, when long
stretches of the embankment structure supporting the tracks are traversing through soft alluvium deposits, cost effective designs meeting the
design performance and construction schedule are required. A cost effective treatment such as PVD with temporary surcharging was
designed to meet the stringent performance requirements of differential settlement of not more than10 mm over a chord length of 10m and a
total settlement of not more 25mm within 6 months after completion. In order to meet the tight construction schedule, basal reinforcement
was also used to allow higher embankments to be built without compromising on the embankment stability during construction. Therefore, in
order to verify the design philosophy of the ground treatment method adopted, a fully instrumented trial embankment was carried out at
Tokai, State of Kedah, Malaysia. This paper presents the consolidation settlement behavior of the trial embankment during construction and
waiting period. Back analyses using Finite Element Modeling (FEM) was performed to evaluate the performance of the ground treatment and
to verify the subsoil parameters used in the design.
Keywords: Trial Embankment, instrumentation, soft ground, consolidation settlement, excess pore water pressure, lateral displacement

1. INTRODUCTION Underlying is dense silty sand to sand from depth of 15m to 24m
depth. Hard layer with SPTN’ value of more than 50 was found
The construction of the electrified double track project for northern below 24m. The general subsoil properties including bulk density,
part of Peninsular Malaysia commenced in year 2007. As most of compression ratio (CR), re-compression ratio (RR), over
the embankments supporting the track are founded on soft alluvium consolidation ratio (OCR), pre-consolidation pressure (Pc),
deposit, cost effective ground treatment such as PVD with undrained shear strength (su) and Atterberg limit are plotted in
temporary surcharging was widely used to meet the stringent Figure 2. The interpreted subsoil parameters based on the field and
settlement requirements. In order to verify the design philosophy of laboratory tests are summarized in Table 1.
the ground treatment method adopted, a fully instrumented trial
embankment was constructed at Tokai, State of Kedah as shown in
Figure 1. This is to study the consolidation settlement behavior, 3. GROUND TREATMENT AND INSTRUMENTATION
excess pore water pressure response, and lateral displacement at the The general ground treatment details for trial embankment are
toe of embankment as indicator of embankment stability during summarised in Table 2. The instrumentation scheme includes
construction filling and rest period. Figure 1 shows an overview of settlement gauges, extensometers, inclinometers, ground
instrumented trial embankment during the surcharging period. displacement markers, vibrating wire piezometers, standpipe and
This paper presents the settlement behaviour of the trial surface settlement markers. Figures 3 and 4 show the cross section
embankment. Back analyses using Finite Element Modelling (FEM) and layout plan of instrumentation scheme. Settlement at centre of
was also performed to evaluate the performance of the ground embankment was measured by settlement gauges SG2, SG5 and
treatment and to verify the subsoil parameters used in the design. SG8. Whilst, settlements at edge of embankment ware measured by
settlement gauges SG1, SG3, SG4, SG6, SG7 and SG9. Settlements
at various depths were measured by extensometers EXT1, EXT2
2. SUBSOIL CONDITION
and EXT3.
The subsoil condition at trial embankment is relatively homogenous
consisting of very soft to soft clay within a depth of 15m.

Human size

Location of trial embankment

Quaternary Sediments
Undifferentiated
Figure 1 Location and overview of instrumented trial embankment
2P-051/ pp. 1
 th
The 17 Southeast Asian Geotechnical Conference
Taipei, Taiwan, May 10~13, 2010
Compression Re-compression Over Consolidation Preconsolidation Undrained Shear Atterberg Limit with
Bulk Density Ratio (CR) Ratio (RR) Ratio (OCR) Pressure (PC) Strength (Su) Water Content

0
1
2
3
4
5
Depth (m)

6
7
8
9
10
11
12
13
14

10 15 20 0.1 0.2 0.3 0 0.02 0.04 0 2 4 6 8 10 0 40 80 120 0 10 20 30 40 0 40 80 120 160

kN/m3 ----- Piezocone kPa kPa %
CR = Cc / 1+eo RR = Cr / 1+eo M - Undisturbed { - Plastic Limit
= - Remoulded Y - Water Content
y - Liquid Limit
Figure 2 Subsoil properties
Table 1 Interpreted subsoil parameters
Depth Soil type SPT'N γbulk (kN/m3) c' (kPa) φ' (o) CR RR OCR su (kPa)
0m to 5m CLAY 0-1 13 5 21 0.25 0.03 3.0 - 4.4 15 - 25
5m to 10m CLAY 0-1 13 5 21 0.22 0.027 1.7 - 2.7 25 - 35
10m t o15m CLAY 1-4 16.5 5 21 0.12 0.017 1.2 30 - 35
15m to 24m silty SAND 12 - 21 18 5 30
24m to 30m silty SAND > 50 18
Table 2 Design Criteria and ground treatment details effect of both the vertical permeability of natural subsoil and radial
Design criteria: Maximum settlement of 25mm and consolidation by PVD was established to simulate the PVD
maximum differential settlement of 10mm behaviour in the back analyses. Based on the back analyses results,
for a chord of 10m (1:1000) over 6 months. kve is about 5.8 times more permeable than the vertical permeability
Soil replacement: 1m deep of the original subsoil (soft clay). The geometry of FEM and the
PVD details: 15m long with 1.2m spacing in triangle input parameters are shown in Figure 5 and Table 3 respectively.
pattern. The back analyses were carried out adopting undrained condition
Sand blanket: 500mm thickness (with generation of excess pore water pressure) with coupled
Basal Geotextile with ultimate tensile strength of consolidation for each stage of construction.
reinforcement: 200kN/m
Surcharge: 1.5m thickness
Stages loading*: 1. 3.9m thick fill and rest for 4 months.
2. 5.8m thick fill and rest for 3 months.
3. 7.6m thick fill and rest for 3 months.
* The multistage construction (with higher height) was carried out
due to site condition and problems such as delay due to wet
monsoon season, no borrow source, etc. The original intent is to
construct the trial embankment in single stage loading of up to 5.9m.

4. BACK ANALYSIS BY FEM MODELLING

Back analyses were carried out by using finite element modelling
(FEM) software (Plaxis). Soft Soil Model (SSM) was adopted to
simulate the behaviour of the soft clay. The SSM is based on the
Modified Cam-Clay Model. Stress dependent stiffness (logarithmic
compression behaviour) between volumetric strain and mean
effective stress is assumed. Distinction between primary loading and
unloading-reloading stiffness based on the modified index λ*
(CR/2.3) and κ* (2RR/2.3) were obtained from 1D-Odometers tests.
In addition, SSM can also memorise the pre-consolidation stress
with OCR input in the initial stage. Hardening Soil Model (HSM)
was utilised to model the underlying silty sand layer and the fill
materials.
From a macro point of view, PVD increases the subsoil mass
permeability in vertical direction (Lin et al, 2006). Therefore, an
equivalent vertical permeability, kve, approximately represents the
Figure 3 Cross section of instrumentation
2P-051/ pp. 1
 th
The 17 Southeast Asian Geotechnical Conference
Taipei, Taiwan, May 10~13, 2010

Figure 4 Instrumentation layout plan

5. MEASUREMENT VERSUS CALCULATION Unfortunately, the extensometers were damage after the first stage
of filling. This is likely cause by the large settlement.
The calculated results of the FEM analyses are compared with the
measured data vertical deformation (settlement). Figure 6 shows the
settlement profile of the embankment at the centre and the edge
versus embankment filling time.
The measured settlements at the end of surcharging period are
averagely 1963mm at the centre of the embankment (which is about
26% of the total constructed embankment height) and 1545mm at
the edge of the embankment. The calculated settlement at the centre
of embankment is 1932mm which is 31mm or 1.6% lowers than the
measured value. Whilst, the prediction on edge settlement is
1457mm which is 88mm or 5.7% lower than the measured value. In
general, the back-calculated settlement profile is fairly close to the
measured settlement profile especially during first stage of filling
(within 200 days) up to a fill thickness of 3.9m.
Settlements at various depths were obtained from extensometer
installed at centre of trial embankment. Figure 7 shows the
settlement profile of the embankment at various depths versus
embankment filling time and thickness of fill.
The extensometers indicated that average settlements at depth of
1.32m, 1.88m, 3.33m, 4.35m and 6.34m were 2084mm, 1828mm,
1672mm, 1351mm and 1108mm respectively. Generally, the back-
calculated settlements in various depths are about 10% to 26%
(186mm to 329mm) less than the measured settlement. The
measured settlement for the top 6m was close to the surface
settlement. The soft clay layer at the depth of first 6m experience
largest portion of settlement as expected. Extensometers were also Figure 5 Geometry of FEM
installed at deeper depth to study the settlement profile with depth.

2P-051/ pp. 1
 th
The 17 Southeast Asian Geotechnical Conference
Taipei, Taiwan, May 10~13, 2010
Table 3 FEM input parameters
Material Soil Model γsat (kN/m3) kx (m/s) ky (m/s) λ* κ* E50, Eoed Eur c' φ'
Layer 1 (0 - 5) SSM 13 2.0E-08 1.0E-08 0.109 0.026 5 21
Layer 2 (5 - 10) SSM 13 2.0E-08 1.0E-08 0.096 0.023 5 21
Layer 3 (10 - 15) SSM 16.5 2.0E-08 1.0E-08 0.052 0.015 5 21
Layer 4 (15 - 24) HSM 18 1.0E-06 1.0E-06 34000 102000 5 30
Layer 5 (24 -40) HSM 18 1.0E-05 1.0E-05 34000 102000 5 30
PVD of Layer 1 SSM 13 2.0E-08 5.8E-08 0.109 0.026 5 21
PVD of Layer 2 SSM 13 2.0E-08 5.8E-08 0.096 0.023 5 21
PVD of Layer 3 SSM 16.5 2.0E-08 5.8E-08 0.052 0.015 5 21
Sand blanket HSM 19 2.3E-05 2.3E-05 30000 90000 0 30
Fill materials HSM 19 - 20 1.0E-06 1.0E-06 30000 90000 5 28

Time (day) Time (day)

0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450

10 10

Fill Thickness (m)

8
Fill Thickness (m)

8
6
6
4

4 2

2 0

Settlement (mm)
-500
0
-1000

-1500

-500 -2000
Settlement (mm)

-2500 At depth of 1.32m

0
Settlement (mm)

-1000 -500

-1000

-1500
-1500
-2000

-2500 At depth of 1.88m

0
At centre of embankment
Settlement (mm)

-2000
0 -500

-1000

-1500
-500
Settlement (mm)

-2000

-2500 At depth of 3.33m

0
Settlement (mm)

-1000
-500

-1000

-1500 -1500

-2000

At edge of embankment -2500 At depth of 4.35m

0
-2000
Settlement (mm)

Measurement Calculation -500

-1000
Figure 6 Settlement of embankment -1500

-2000
6. CONCLUSIONS At depth of 6.34m
-2500

Based on observations of the trial embankment performance and the Measurement Calculation

analyses results, the following conclusions are made: Figure 7 Settlement of embankment at various depths
a) The total settlement at the end of surcharge period is about
26% of the constructed embankment height.
b) The measured settlement at original ground level was about 7. REFERENCES
1.6% to 5.7% (31mm to 88mm) more than the calculated
[1] D.G.Lin, W.T.Liu and P.C.Lin, Journal of the Southeast
settlement.
Asian Geotechnical Society, December 2006, pg. 157 – 170.
c) In general, the settlement was reducing with depth under the
embankment. The calculated settlements in various depths
were about 10% to 26% (186mm to 329mm) less than the
measured settlement.
d) In finite element modelling, an equivalent vertical
permeability, kve, approximately represents the effect of both
the vertical permeability of natural subsoil and radial
consolidation by PVD can be adopted to simulate the PVD
behaviour.
e) Back analysis by using equivalent vertical permeability
method for PVD treatment is about 5.8 times more permeable
than the original soil permeability.
f) The settlement measured for the first stage filling up to 3.9m
has good agreement with the settlement computed using SSM
with coupled consolidation and undrained
behaviour. However, the settlement measured and computed
at the end of surcharging period differ by less than 6%.

2P-051/ pp. 1