Twelf Southeast Asian Geoteclmical Conference, 6-1 O May 1996, Kuala Lumpur 199

T. S. CHUA, B. C. TAN and T.L. YU

L&M Geotechnic (Singapore) Pte Ltd

Deep Excavation Work Near MRT Structures

SYNOPSIS The convenience of the fast and reliable Singapore MRT system attracts many building development projects
near its stations. However, the construction activities especially geotechnical work would require special consideration to
minimise any disturbance to this existing sensitive structures. This paper briefly describes a deep excavation work using a
unique semi Top-Down method for a major development project near a MRT station and viaduct. The design of the
diaphragm wall which acts a temporary retaining wall as well as a permanent wall, using a 1-D and a 2-D analysis will be
discussed. The result of an extensive inst1umentation scheme used to monitor the effect of the excavation work on the
smTOunding will be reported and compared with the predicted values.

1.0 INTRODUCTION 2.0 DESCRIPTION OF PROJECT

The fast and reliable Singapore Mass Rapid Transit (refer to As shown in Fig. 2, the project is located beside Boon Lay
Fig. 1), which began construction in late 1983, has been in MRT Station and is bounded by Jurong West Central 2 and
operation since late 1987. The convenience of the Upper Jurong Road. It involves the construction of 3-level
efficiently run Mass Rapid Transit attracts many building basement for carpark and shopping and 6 storeys of
development projects nears its stations. The construction shopping and offices. Diaphragm wall was proposed as
activities, however, have to be carried out in a manner that earth retaining structure during const'ruction of basement
will not cause any significant disturbance to these existing and form part of permanent structure after the co1npletion of
sensitive structures. Under this condition, conventional the building. The lowest excavation level is 12.0m below
method of analysis and design may not be adequate, as the ground level. The Boon Lay MRT Station is an above­
magnitude of movement of the surrounding structure ground structure supported on bored piles foundation.
caused by the works will govern the design. This paper
aims to· highlight some of the design consideration,
techniques and construction method for an excavation work
for a prestigious project near a MRT station.

=
lll\D U�

Q
lli:'<'.TtD

Q ��!Al"'!<
"' o:n= ""'"'�

0 0.t\'.<.1<I> ..AUON

Fig. l Singapore Mass Rapid Transit System

 T. S. CHUA, B. C. TAN and T. L. YU
L&M Geotechnic (Singapore) Pte Ltd

3.0 GEOLOGICAL INFORMATION 5.0 METHOD OF CONSTRUCTION

The subsurface soil profile on site consists of firm silty clay In order to meet the completion date specified by the
fill followed by swampy soils of very soft peaty clay. This developer, a unique method of construction was adopted.
As shown in Fig. 3, during the initial excavation stages, two
is underlain by residual soils of firm to hard clayey silt. The
residual soil layers are founded on bedrocks of siltstone levels of composite struts were used to support the
and sandstone (Jurong Formation). The thickness of the fill diaphragm wall during the bulk excavation to the B2
varies from l .3m at the eastern end to 4.2m at the western basement slab level. The second level of strut were
removed when B2 basement slab has gained sufficient
end of the site. Similarly, the very soft peaty clay layer
strength. This is followed by the casting of BI basement
becomes thicker in the westward direction. The thickness of
slab and removal of first level strut. The first storey slab
residual soils ranges from 6.1 m to 2 l .3m. The bedrock
was then cast to provide sufficient rigidity before
formation appears to dip towards the south from the
excavation proceed in a Top-Down manner from B2 level
northern end and eastern end of the site.
to B3 level. At the same time, construction for the
superstructture was carried out. In order to facilitate Top­
Down construction from B2 level to B3 level, plunge-in
4.0 THE DIAPHRAGM WALL SYSTEM
column using steel H-section were installed. The plunge-in
column shown in Fig. 4 were designed to take part of the
In view of the proximity of the MRT station, a stringently
superstructure load during construction stage and were later
designed retaining wall system is required to ensure safe
incorporated into the permanent column to carry full service
excavation works. The presence of bedrock at relatively
load. The plunge-in column were installed, using a steel
shallow depth rules out the use of sheetpiles. Prebored H­
frame guide, coincide with bored piles which is the
piles was not feasible due to potential ground water
foundation system for this project.
drawdown problem. After studying various methods, it was
found that diaphragm wall was the most feasible solution.
The diaphragm wall acts as earth retaining wall during 6.0 METHODS OF ANALYSIS AND DESIGN
excavation works (temporary stage) and formed the
basement wall of the substructure. The layout of the Typical sections of the diaphragm wall are identified
diaphragm wall is detailed in Fig. 2. The diaphragm wall according to the proximity of the MRT station, depth of
system consists of 800mm thick panels near MRT reserve excavation and subsoil conditions. A one dimensional finite
and 600mm thick panels for area away from MRT reserve. element computer program was used for the analysis of the
retaining wall. The program has the capability to model the
interaction between the soil, retaining wall and struts taking
into account the sequence the construction and the non­
UJnW
reversible stress-strain characteristics of the soil. The soil is
JMOOnC Wt!! >

.,;.
--,tr--==;;c--=;'::;: :::�
i<�:-.,� ::;: modeled based on the Rankine approach (elasto-plastic soil
.. � . ...,.,, ....
<>-•"""'·"'

·r,:>:.��'li" behaviour) with provision for soil-wall friction.

T ,.,...... ...,......

·- .....�""'"

�J
.. ::.=�.;,:1Hl
...

�:· ,,.

: ......� .. .
VXllT SQIT
�1.ATY Cl.AT

• <

�
.. �
•,

c
"'1".

aDi: =-·"'"'-

!' ·

Sll,TSTO�f. IQO

Dl1phrtgm Woll for £Juv1t1o� Worlr.o Neu Mll.T Structur<3

Fig. 2 Site Layout Plan and Instrumentation Fig. 3 Typical Section

 T. S. CHUA, B. C. TAN and T. L. YU
L&M Geotechnic (Singapore) Pte Ltd

7.0 EFFECTS OF DIAPHRAGM WALL

CONSTRUCTION AND BULK EXCAVATION ON
MRT STRUCTURES

As part of the excavation works fall within the MRT

reserve, it is a requirement to study the effect on the
existing MRT structure due to diaphragm wall construction
and bulk excavation works.

The main concern regarding diaphragm wall construction

was the stability of the slurry trench. The present of the
peaty clay with very low strength Vfas initially thought
Fig. 4 Typical Plunge-in
could be prone to collapse. Trench stability analysis shows
Column Detail that l m slurry head of Bentonite above the existing ground
level was required. The effect of the diaphragm wall
'
construction was carried out by field trial and monitor the
The undrained shear strength (C,) is computed based on 4N works using instrumentation. A diaphragm wall panel
where N is the SPT value. The coefficient of the subgrade located outside MRT reserve was selected. During
reaction is assu1ned to be 300 times the Cu values. installation, the soil movement was monitor using
inclinometer preinstalled at a distance l m away.
The stages of excavation and basement construction for 3- Inclinometer reading was taken at regular interval during
level basement simulated using the computer analysis are trenching. The results indicated that the soil movement was
described as follows (refer to Figure 3) : insignificant.

Stage 0 Installation of diaphragm wall The effect of the bulk excavation on the MRT structure was
analyzed using a two dimensional finite difference analysis.
Stage 1 Excavation to 3.2m below !st storey level The soil was based on the Mohr-Coulomb failure criteria,
Stage 2 Installation of !st row of struts at 2.7m while the diaphragm wall, MRT piles, slabs and struts were
modeled as beam elements. Before modeling the actual
Stage 3 Excavation of 6.4m level construction activities , a finite difference mesh was set up
Stage 4 Installation of 2nd row of struts at 5.9m and boundary conditions imposed . The size of the mesh is
set up to be large enough so that the effect of boundary
Stage 5 Excavation to 7.5m level conditions on the analyzed zones may be minimized.
Stage 6 Casting of basement 2 slab at 7.0m level Initially undrained soil parameters were adopted for total
stress analysis and gravity run was carried out to simulate
Stage 7 Removal of 2nd row of struts the existing soil conditions, followed by modelling the
Stage 8 Casting of basement I slab at 3.5m level installation of MRT piles and diaphragm wall. The results
of the analysis (Fig. 5), suggested that the effect on the
Stage 9 Removal of 1st row of struts existing MRT pile is not significant.
Stage I 0 Casting of 1st storey slab
Stage 11 Backfill soil to I st storey level -�

Stage 1 2 Excavation to formation level at 1 2.0m level

Stage 1 3 Casting of basement 3 slab
Stage 1 4 Long term analysis using drained parameters. 01/.Pl!RA!;M WAU
mo• H.�lmm

Based on the analysis, the required diaphragm wall depth MFH Pll£S
mo>.1.S2nvn
was 20m. The maximum bending moment for 800mm and
600mm wall were 933 kNm/m and 773 kNm/m,
respectively. The estimated strut load were 15T/m for the
first level and 20T/m for the second level. The required
preload was 6.25T/m for both levels.
Fig. 5 Results of 2-D Analysis
 T. S. CHUA, B. C. TAN and T. L. YU
If
L&M Geotechnic (Singapore) Pte Ltd

8. GEOTECHNICAL INSTRUMENTATION

In view of the existing MRT structure, an extensive

instrumentation program was implemented on site. The
program was used to verify the performance of the retaining
wall to ensure that it was acceptable and in agreement with
the predicted behaviour.

Nine clusters of monitoring instruments, each consisting of �·"- -

::o:::.
an inclinometer, a piezometers and a water standpipe were - ""

located at strategic locations . The location of these

instrumentation are given in Fig. 2. Inclinometers (2 nos )
� ..... ,.�......
were also installed in the diaphragm wall fronting the MRT · --....... .....
. ... ...,, ,,_

structures. In addition, tilt plates, vibration sensors and

survey prisms were installed on nearby MRT viaducts
shown in Fig. 6, to monitor the response of the structures. Fig. 6 Instrumentation For MRT Structures
During the excavation and construction of basement
structure, instrument readings were taken twice weekly.
After the completion of basement structure, the frequency
of reading taking was reduced to once a month.
- -· ·--·--·-- _,_

!
0 0 0
!
I
I
2
-
- 2 1-� I i
·--
2
i
!
i

'

4 4 4 -

6 6 6

8 8 f--- ·-
I- 8
T
. I I
1--,�i_Jl
10 10 10

E 12
I/ I E 12
I

E 12
I I
.c .c
I
r
a.
.,
a.
., -4./ %
.,

1 '/
14 14 0 14
0 0

'!I ,1 I
'II I
16 16 16
I

./ I I
I
7/ I I
18 18 18
I
• Stage 1 • Stage 1
20 f--*l'----H o Stage 3 20 o Stage 3 20
I Final Stage
,. Stage 5 ;. Stage 5 i • 13A
--- --
22 - 22 22 --·- -�··-

'
o Stage 12 o Stage 12 • 14A

+Final +Final 0 1D
24 24 24

I I
D 2D

26 -- I I J
26 26
·5 0 5 10 15 20 25 ·3 0 3 6 9 12 15 .5 0 5 10 15 20 25

Deflection (mm) Deflection (mm) Deflection (mm)

Fig. 7 lnclimometer 13A Readings Fig. 8 Inclinometer 14A Readings Fig. 9 ·co111parison of Lateral Movement
'

T. S. CHUA, B. C. TAN and T. L. YU

L&M Geotechnic (Singapore) Pte Ltd

Inclinometers, BA and 14A, were installed in diaphragm

wall next to MRT structures. The deflection of the wall at
various stages of construction are shown in Fig. 7 and Fig.
�--�I
8, respectively. The maximum deflections of the wall at ' '
'
final stage were approximately 16 mm at BA and 8 mm at
14A. 14A being nearer to a corner (3-D effects) recorded
half the values that of BA. The predicted deflection curves
at final stages from 1-D and 2-D computer analysis and the
measured values are presented in Fig. 9. The 2-D analysis .-..;oITTomr
;•MTER�OON
predicted the maximum deflection near ground surface l.m..i.YmS
quite accurately as compared to 13A. However, the bPnECOA!l
deflection below ground was under estimated. The results I
of the 1-D analysis was smaller and was close to the • ,J ___ . .j.I • !
Altcrlnstollo!ion of2nd •trul
measured deflection at 14A. It was also observed that, the
measured movement registered by inclinometer installed in

·
soil immediately next to the wall was generally smaller. For
example, the maximum movement in soil recorded in I', 1 '
II

" LLI..LLl-'.J.C..L..J��'=-'-:;;;:;-�.LI
I
inclinometer, 13, was 10 mm as compared to 1 6mm in the
"'"'°1:-10-�··0-E:.,,.':''·t.••.�··:..,.'it''fo.,,�···�... �··-r..,,...
later. ��

The strut loads were measured by the load cells attached to

the jacking unit and were generally within the design
values. Fig.10 shows a typical result (strut at level 1). It
can be noted that the strut load may fluctuates due to Fig. 1 0 Typical Strut load (for level I)
temperatures effects. It was also observed that the strut
load first increased above the preload following further
excaVation, and then drop when the next strut layer was
installed and preloaded.
9.0 CONCLUSION

The measured water table and piezometric levels fell by 1m

to 2m closet to the MRT line and is more (up to 4m) at the The basement structure of the development was
not1hem end due to shallower diaphragm wall. The changes successfully constructed using the diaphragm wall and semi
in the water level are caused by a combination of factors, Top-Down construction sequence. With the unique
namely seasonal variation due to very dry weather during construction 1nethod, the project was con1plete within
the construction period, effect of unloading due to soil schedule. With careful planning, design and supervision,
re1noval and some seepage through the tips of the limiting the movement of the sun·ounding structure to a
diaphragm wall. The survey on surface settlement markers very small magnitude is achievable.
showed that the surrounding ground settled between 15mm
to 64mm.
ACKNOWLEDGEMENT
The tilt plates installed on the columns of MRT viaducts
registered showed readings of between 4 seconds to 33 The authors are grateful to Jurong Point Reality Pte Ltd for
seconds. The movements of columns monitored by survey permission to publish this paper. The authors also wish to
prism was less than 5mm were observed. The vibration express their gratitude to Land Lease (Singapore) Pte Ltd
level due to site activities was monitored using vibration for permission and providing much of the information used
sensors. The measured vibration level on the MRT in the preparation of this paper. The structural consultant is
structures was below 8 mm/sec which is less than the limit Maunsell (Singapore) . Special thanks are due to Mr.
of l 5mm/sec. William Them for his kind assistance.
View publication stats