Underground Singapore 2009

Assessment of CCL C824 Boulevard Siding Earth

Retaining System and Proposal for Strengthening Works
K. Y. Yong, W. M. Tsui, H. G. Tay & C. Y. Cheng
AECOM Singapore Pte Ltd (formerly known as Maunsell Consultants (S) Pte Ltd)

C. C. Ng
Land Transport Authority, Singapore

T. Furusono, Y. Ueda
Nishimatsu Construction Co. Pte Ltd, Singapore

ABSTRACT: Cut and cover excavation was carried out for Circle Line Contract 824 Boulevard Siding
over a length of 532m, 20.5m deep and 24.0m wide to construct twin box tunnels in thick Kallang
Formation soil underlain by Old Alluvium. Large diaphragm wall deflections were recorded when ex-
cavation reached the 4th to 5th strut level. Despite implementing mitigation measures, excessive wall
deflection persisted down to final excavation level until the base slab was cast. This paper describes
the assessment of the temporary cofferdam and strengthening works design to allow for work to re-
sume in a safe environment. The diaphragm walls were assessed through curve fitting by inclinometer
readings and Finite Element (FE) modeling of the excavation stages. Predicted zones of yielding were
consistent with the observation of large cracks that developed on the diaphragm wall. Safety enhance-
ment steel strutting and bracing system was carefully designed and installed over the areas with dis-
tressed diaphragm wall. Construction over the affected area has since been completed successfully.

1 INTRODUCTION

Circle Line Contract C824 involved the construction of 2 stations and tunnels running from Millennia
to Mountbatten Road as show in Figure 1. The tunnels were to be constructed using cut and cover me-
thod with the exception of the alignment under the Kallang Basin where bored tunneling was em-
ployed. Towards the east of the Kallang Basin, the cut and cover tunnels forming Boulevard Siding
(BLS) were designed to be retained by 600mm thick temporary diaphragm walls embedded into Old
Alluvium (OA) with 5 to 7 levels of preloaded struts. Excavation for this area was to proceed in chal-
lenging soft ground conditions with up to 34m deep of soft marine clay.

Figure 1. Location of Boulevard Siding with respect to Circle Line Contract C824

220
 Underground Singapore 2009

During excavation to 4th and 5th level strut, at about 14m below ground level, wall movements in-
creased significantly over the eastern most third of BLS alignment. Maximum recorded inclinometer
movement was approximately 500mm at Type W2a bracing system. From safety considerations, miti-
gation measures which included jet grout pile (JGP) ground improvement, additional struts and moni-
toring instruments were introduced. However, when excavation resumed in October 2003, wall deflec-
tion persisted at most areas up to a maximum of 700mm. As an urgent risk mitigation measure, the
casting of base slab for permanent structure was expedited to stabilize the situation. Figures 2 and 3
show the horizontal cracks on the diaphragm wall and deformed bracing system due to excessive wall
movement. Figure 4 presents the rapid development of inclinometer movements adjacent to the South
retaining walls with excavation stages.

Due to the large wall movements observed on site, work was suspended along the BLS excavation.
Further construction of the tunnel structure was to commence after the condition of the diaphragm wall
has been assessed and the safety condition of the temporary cofferdam improved, where necessary, to
the acceptance of various authorities.

In this paper, the terms current or existing cofferdam refers to the condition of the cofferdam in June
2005 at the start of the back-analysis and design of the strengthening works.

STRUT S6

KINGPOST
OUT OF
VERTICALITY VERTICAL
HORIZONTAL CRACKS
ALONG DWALL

STRUT S7

Figure 2. Observed cracks on diaphragm wall Figure 3. Existing shoring condition

800

700 FEL/Cast Base Slab

South Wall Max Inclino Disp (mm)

600 Exc S6 FEL/Cast Base Slab

500

400
Exc S5
Exc S7
300
Exc S6
200 Exc S4 I79 - W2a
I84 - W1b
100
I83 - W1a

0
Exc S1 Exc S4
20-Mar-03 28-Jun-03 6-Oct-03 14-Jan-04 23-Apr-04 1-Aug-04 9-Nov-04 17-Feb-05 28-May-05
Date
Figure 4. Development of inclinometer displacement (adjacent to dwall) with excavation stages

221
Underground Singapore 2009

2 GROUND CONDITIONS

The affected length of tunnel alignment is located within thick Kallang Formation deposits comprised
predominantly of normally to slightly over-consolidated marine clay. The soft marine clay is encoun-
tered down to a depth of 34m below ground level and is underlain by competent Old Alluvium mate-
rial. Figure 5 indicates the contours of OA surface along Type V to Y bracing system. The OA is ob-
served to dip from the north towards the south diaphragm wall. This difference in OA levels on both
sides of the cofferdam is prominent over Type W2/W2a/Y and has significant impact on the wall de-
flections due to unbalanced loading conditions as presented in the subsequent sections.

Figure 5. Old Alluvium contours along cut and cover tunnel alignment

3 ASSESSMENT OF TEMPORARY EARTH RETAINING SYSTEMS (TERS)

3.1 Diaphragm Wall

Curve fitting was performed based on the inclinometer readings to assess the current wall condition.
Curvature values at critical locations (excavation side) were identified and corresponding cracked sec-
ond moment of inertia (Icrack) was calculated to derive wall bending moments. The assessment for Type
W2a South wall concluded that a plastic hinge is likely to have developed near to S6 strut level from
excavation to strut S5 onwards as shown in Figure 6a.

Numerical back analysis was also performed to compare the findings obtained from the curve fitting
exercise and to subsequently assess the condition of the wall during strut removal stages. The follow-
ing assumptions were made in the back analysis to more realistically model and simulate the complex
excavation sequence up to casting of base slab:

222
 Underground Singapore 2009

x C824 design soil parameter table as shown in Table 1

x Surcharge excluded to realistically simulate site conditions
x Rebar yield stress of 500N/mm2 (based on mill certificates) and material factor Jm of 1.0
to determine wall moment capacity
x Load factor of unity to assess the wall under working load conditions
x Flexural stiffness (EI) reduced from 80% of EIgross to 50% of EIgross over yielded portion
of the wall to account for equivalent cracked section
x Moment capacity of wall defined in analysis to allow for development of plastic hinge

Table 1. Back-analysis soil parameters

Stratum Bulk Density Cu c' I' E k Ko

(kN/m 3) (kPa) (kPa) (deg) (kN/m 2) (m/s)
Fill 19 25 0 30 10000 1x10 -6 0.5
10 (0-5m) 4000
M 22 (Upper)
16 20 (5-15m) 0 8000 1x10 -9 1.0
(Marine Clay) 24 (Lower)
20+1.6(z-15) 400C u
F2 20 (0-10m) 24 6000
19 0 1x10 -9 1.0
(Fluvial Clay) 20+2(z-10) 400C u
F1
20 0 0 30 1500N 1x10 -6 0.7
(Fluvial Sand)
E 15 (0-10m) 18 6000
15 0 1x10 -9 1.0
(Estuarine Clay) 15+2.3(z-10) 400C u
OA N<30 20 0 (10)* 32 (20)* 5x10 -7 1.0
OA 30<N<50 20 5 (20)* 32 (25)* 5x10 -7 1.0
5N 2000N
OA 50<N<100 20 5 33 5x10 -8 1.0
OA N>100 20 10 35 5x10 -8 1.0
JGP 16 300 - - 150000 1x10 -10 0.7

* values are for clayey material of Old Alluvium Formation

z - depth below ground level

105 105

SOUTH S1 S1 NORTH
100 100

S2 S2

Exc to S2
95
S3 S3 95
Exc to S3

S4 S4
Exc to S4
90 90
S5 S5

S6
RL (m)

S6
RL (m)

YIELDED
85 85

FEL FEL
Exc to S5
80 Exc to S6 80
Exc to FEL
Exc to S1
Exc to S1
Exc to S2
Exc to S3 Exc to S2
75 75
YIELDED Exc to S4 Exc to S3
Exc to S5 Exc to S4
Exc to S6 Exc to S5
Exc to FEL Exc to S6
70 Cast Base Slab Exc to FEL 70
Mult fy = 500
Cast Base Slab
Mult fy = 500
Mult fy = 500
Curve Fitting
Mult fy = 500
65 65
-2000 -1000 0 1000 2000 3000 4000 5000 6000 7000 -4000 -3000 -2000 -1000 0 1000 2000
BM (kN.m) BM (kN.m)

6(a) 6(b)

Figure 6a and 6b. Back-analysed bending moment along south and north wall using curve fitting
and finite element methods

223
Underground Singapore 2009

Analysis results generally indicate that the South diaphragm walls have yielded at two distinct loca-
tions; on the soil side at the level of OA-Kallang Formation interface and on the excavation side over
the area close to formation level. As the base slab has been cast, yielding of wall below formation level
on the soil side is expected to have minimal effect of the stability of the cofferdam during strut re-
moval stages. Assessment of the North diaphragm walls generally indicate minimal or no yielding
with the exception of Type W1b retaining system where thick marine clay (down to 33mBGL) was
encountered.

The development of bending moments and yield zones over the Type W2a diaphragm walls is pre-
sented in Figures 6a and 6b. The predicted area of yielding on the South wall is generally consistent
with the results from the curve fitting assessment and location of horizontal cracks observed on site.
Figure 7 presents a typical analysis model for Type W2a cofferdam system.

It is noted that emphasis was placed on observing the progressive development of diaphragm wall be-
haviour in terms of trends rather than attempting to match wall deflections with monitored results. This
is due to the fact that the ground conditions were highly variable and that the cofferdam system has
been subjected to complex activities rendering it sensitive to slight changes in external factors.

24.9m

RL103.0
102.0
Fill
98.0
Fill 98.6 97.9
95.6 F2 97.9
92.6
M M
89.1
96.8
86.0 86.1
83.9 F2 OA N~13 87.5
82.5 (FEL)
OA N~21 87.5
JGP
M OA N~41 86.3
RL76.9
75.7
OA N~80
F2 84.0
68.9
66.9 OA N~21
RL66.0
62.9 OA N~80 OA N>100
OA N>100

Figure 7. Finite element analysis model at Type W2a earth retaining system

3.2 Steel Bracing System

Diaphragm wall panels were installed in 4m widths while struts were spaced at 3m c/c resulting in an
alternating series of 1 and 2 struts per panel. As the walers were not continuous over the wall panels,
struts were assessed based on a maximum tributary load length of 4m (forces from numerical analysis)
conservatively assuming no shear transfer between the panels. Results indicate that the back-analysed
strut forces are generally within ultimate capacities.

Strut lift-off test was subsequently performed to determine the existing loads in the current condition.
The measured forces were generally less when compared with the numerical analysis predictions
which could be due to redistribution of loads from the stage of casting base slab to the current condi-
tion where creep of the soil and the structural retaining system could have taken place.

224
 Underground Singapore 2009

3.3 Failure Mechanism Analysis

Assessment of the diaphragm wall and bracing members during excavation to formation level has
demonstrated that the current condition was stable and in equilibrium. However, considering the dis-
tressed state of the diaphragm walls, failure mechanism checks were further performed to ensure con-
ditions within the cofferdam were safe to resume construction works. Two checks were carried out; re-
consolidation of soils to at rest (Ko) pressures and one strut failure scenarios with concept as shown in
Figures 8 (a) and (b). Reconsolidation of soil to Ko conditions was performed as monitoring results in-
dicate that the walls were creeping laterally (max 30mm) since casting of base slab and that works
within the excavation is expected to resume more than 1 year after completion of base slab.

As wall failure requires 3 hinges to be formed over any supported span, the above checks conserva-
tively assume that a full plastic hinge has developed on the excavation side of the wall within the vi-
cinity of formation level. Remaining 2 hinges are required to develop on the soil side for failure to be
possible. Results indicate that induced diaphragm wall forces were within ultimate capacity and that
failure mechanism would not develop to destabilize the cofferdam under the various possible loading
conditions.

Type W2a Area (BLS)

103.0

103.0

95.0

95.0
Increase Soil
Pressure

RL (m)
by 40% 5th Strut
Overall 2 5th Strut BM
2 6th Strut BM
RL (m)

FOSt 1.4 85.0

1
3 Base BM P6 85.0
1
3 Base BM
Current
Soil
Pressure
FOSt 1.075.0
75.0

65.0 65.0
-2000 -1000 0 1000 2000 3000 -2000 -1000 0 1000 2000 3000
BM (kNm / m) BM (kNm / m)

8(a) 8(b)

Figure 8a and 8b. Reconsolidation to Ko pressure and one strut failure mechanism analysis

4 PROPOSED STRENGTHENING WORKS

Proposed strengthening works to the existing cofferdam system was carried out in two stages.

Stage 1 of the strengthening works includes the installation of a new kingpost and runner beam system
to the existing cofferdam system as the existing kingpost and runner beam was observed to be in bad
condition. The new king post and runner beam system were designed to support both the existing cof-
ferdam system and the new strengthening system.

Stage 2 of the strengthening works involves the installation of a new cofferdam strengthening system
at the last two to three strut levels depending on the results of the numerical back analysis and the fail-
ure mechanism analysis. The new cofferdam strengthening system was installed at areas of the dia-
phragm wall where the wall was found to have yielded based on the results of the numerical back
analysis or where there was a possibility that the wall would yield based on the results of the failure
mechanism analysis. The new cofferdam strengthening system consisted mainly of evenly spaced steel
H-piles placed against the existing diaphragm wall and propped by the wall kicker cast from the tunnel
base slab and the additional struts above. Continuous waler beams were provided for load transfer
from the H-pile wall to the struts.

225
Underground Singapore 2009

The additional H-pile wall above the base slab was non-composite with the existing diaphragm wall
and provides additional stiffness (EI), shear and bending capacity to the existing diaphragm wall,
therefore, enhancing its factor of safety of the strengthened zone. The additional wall elements shares
the additional loads, thereby preventing the formation of additional plastic hinges above the base slab
and ultimately the onset of a failure mechanism The sizing and capacity of these additional H-pile
wall and strutting system were determined based on the required additional capacity to achieve ulti-
mate load factor of 1.4 (Dead + Water + Earth) + 1.6 Imposed Loads from subsequent stages of the
works and ultimate load factor of 1.05 (Dead + Water + Earth + Live) for accidental strut removal.

The new cofferdam strengthening system was constructed with the existing strutting system in place.
Both the struts of the new cofferdam system and that of the existing cofferdam were kept in place until
the tunnel wall is cast to the base of each existing strut level. Re-strutting against completed tunnel
walls and plugging of the gap between the tunnel wall and diaphragm wall was carried out before the
removal of the both the new and the existing struts.

A figure showing a typical cofferdam strengthening system at Type W2a diaphragm wall is provided
in Figure 9.

Figure 9. Typical section for Type W2a strengthening works

5 MONITORING

When construction works for BLS tunnels resumed on 2 January 2005, monitoring of the cofferdam
comprised of 4 piezometers, 17 sets of strain gauges, 9 inclinometers and 78 nos. heave markers on the
base slab. As most of the inclinometers were installed in ground that is consolidating, it was suspected
that the subsequent readings of the ‘lateral ground movement’ recorded could be contributed by the
consolidation component due to compression of inclinometers without telescopic coupler. Hence, 10
sets of wall prisms were added inside the cofferdam in line with the inclinometers to comprehend and
verify the accuracy of inclinometer readings. Others ground monitoring instruments includes ground
settlement markers and crack meters.

From 2 Jan 2005 until 8 September 2007 (backfilling works completed), all instruments registered
readings that are within acceptable limits as the cofferdam experienced small movements less than
10mm during this period. The diaphragm walls displaced a maximum of 6mm during strut removal
stage which is in close to the predicted value while maximum monitored base slab heave was 5mm.

226
 Underground Singapore 2009

6 CONCLUSIONS

Back analysis was performed to assess the condition of the Boulevard Siding cofferdam system based
on the available monitoring readings and site observations. Finite element and curve fitting methods
revealed that the diaphragm walls have yielded at areas close to formation level and within Old Allu-
vium soil resulting in the large observed deformations. Strut forces obtained from FE analysis and lift
off tests indicated that the condition of the existing bracing system was within ultimate capacity but
less than the required factor of 1.4 thus requiring strengthening works. New kingposts, soldier piles
and strutting was installed over the bottom 3 strut levels to enable construction of the box tunnels to
complete. The diaphragm walls performed satisfactorily during strut removal stage with lateral move-
ments being limited to 10mm. The work was completed successfully with the implementation of the
appropriate strengthening works after detailed assessment of the distressed cofferdam retaining sys-
tem.

227