Underground Singapore 2003

A Floating-Type Braced Excavation in Soft Marine Clay

P.C. Lim
CL Geoconsultant, Singapore

T.S. Tan
ECON Corporation, Singapore

ABSTRACT: The behavior of braced excavations for the construction of a services tunnel located in
soft marine clay is described. The total length of the 2-cell tunnel to be constructed using the cut-and-
cover technique in the first phase is 1.5 km, with depth of between 12 and 16 m generally. The ground
condition at the site consists typically of reclaimed fill overlying deep-seated soft, highly plastic ma-
rine clay with depth varying from 20 to 40 m. A floating-type sheet pile cofferdam with a jet-grouted
layer located just below the final excavation level is the proposed temporary ground support method
for the construction of the cut-and-cover tunnel. This paper is a joint effort between the designer for
temporary works and the contractor to present the design considerations and actual behavior of the
braced excavations. Case studies of two excavations: a collapse case and a successfully completed
case are described.

1 INTRODUCTION

The proposed 1.5 km services tunnel is located on Reclaimed land. Depth of the tunnel is typically be-
tween 12 and 16 m. The tunnel is to be constructed using the principle of cut & cover technique. Re-
claimed sand fill overlying very soft to soft marine clay extending to significant depth is a typical
characteristic of the ground condition at this site. Thickness of the soft and highly plastic marine clay
varies from 20 to 40 m but depth of 35 to 40 m is more typical generally.

In such deep-seated soft ground condition and considering the relatively shallow tunnel depth, the
choice of an appropriate yet cost effective temporary ground support method for the construction of
the cut & cover tunnel becomes a critical cost factor for the contractor. The concept of a floating-type
sheet pile earth retaining wall was first envisaged from Mana & Clough (1981). More recently, Weis-
senbach & Kempfert (1995) describes a project located in soft lacustrine clay where the excavation ad-
jacent to an existing building was supported using strutted sheet pile wall, high pressure injection slab
and injection piles. Another example of a cofferdam with grouted bottom forming a water-tight box in
pervious soil condition is also reported by Guilloux & Kastner (1995). Feasibility of the concept was
discussed with B. B. Broms through personal communication initiated by the first author.

A floating-type sheet pile earth retaining wall (cofferdam) with a jet grout layer at the bottom of the
excavation and braced internally with multiple strut layers was then proposed as a viable ground sup-
port method for the construction of the cut-and-cover tunnels. Details of the proposed support method
are described in the next section.

1.1 Objective

The concept and key design considerations of the floating-type temporary ground support are de-
scribed. The primary objective of this paper is to report the actual performance of this less conven-
tional ground support method. Two case studies with different excavation depths are discussed in this
paper.

Main
326
2 BACKGROUND

2.1 Ground condition

The ground conditions typical at the sites of the two case studies designated as Site A and Site B are
shown in Figures 1 and 2, respectively. Reclaimed fill of predominantly beach sand 9 to 10.5 m thick
overlying deep-seated soft marine clay extending to depths varying from 35 to 39 m below the ground
surface is typical. Locally, fluvial clay or/and fluvial sand is present as an intermediate layer dividing
the marine clay into two distinct layers. Below the marine clay lies the Old Alluvium formation. The
depth to the ground water level is generally about 3.0 m below the ground surface at low tide and at
high tide, the water level could rise to about 1.5 m below the ground surface. Summary of the
undrained shear strength of the marine clay obtained from laboratory undrained unconsolidated tri-
axial compression tests (UU) and in-situ field vane shear tests are also given in Figure 1.

Undrained Strength of Marine Clay

cu (KPa)
0 10 20 30 40 50 60 70
0

10
UU
15 FVT

20

25

30

35

45
Design Line

Depth (m)

Fig. 1 Ground condition at Site A and summary of the undrained shear strength of marine clay

Fig. 2 Ground condition at Site B

2.2 Proposed method of temporary ground support

The widths of the cut-and-cover tunnels are generally 18 m for the main tunnel and 16 m for the
branch tunnel. The corresponding width of the excavation is wider by about 3 m typically. Depth to

327
the tunnel soffit is generally between 16 and 16m. Locally at two locations, the shallowest section is
only 9 m below the ground surface while the deepest section is 22 m. The shallowest section is located
at where the tunnel crosses over the existing MRT tunnels while the deepest section is where the tun-
nel crosses below the 8 m wide existing drain.

A schematic of the typical temporary ground support termed herein as the floating sheet pile cofferdam
is illustrated in Figure 3. The proposed ground support is adopted at nearly 90% of the cut-and-cover
tunnels where deep-seated marine clay is present.

Sheet Piles Sheet Piles

Reclaimed Sand Fill

Strut layers
Very Soft to Soft Marine
Clay
JET GROUT LAYER

Medium Dense to Very

Dense to Hard Old
Alluvium

Tension Piles
Fig. 3 A schematic of the floating-type sheet pile earth retaining wall

The temporary ground support consists of interlocking sheet piles with the piles terminating a few me-
tres below the tunnel soffit, forming a cofferdam around the excavation. The sheet pile wall is braced
internally with multi-strut layers above the final excavation level and a jet grout layer located just be-
low the final excavation level. Thickness of the jet grout layer varies from 2.5 to 3.5 m generally and
up to 4 m at the deepest section, depending on the excavation depth and the design requirements spe-
cific to prevailing ground condition.

With the sheet piles floating in the soft marine clay, the primary design consideration is the stability
against basal heave. The stability against base heave is achieved by providing a layer of overlapping
jet grout piles below the final excavation, to form a continuous layer across the excavation. The role of
the jet grout layer is dual purpose. Firstly, it acts as a continuous slab in resisting the bearing pressure
exerted by the adjoining ground and secondly, as a strut in restraining the toe movement of the sheet
piles. The jet grout layer is an un-reinforced structural member and overlapping jet grout piles to form
a homogenous layer providing adequate flexural, shear and compressive strengths is therefore crucial.
The design calls for a required unconfined compressive and tensile strengths of the jet grout columns
of 1200 and 250 kPa, respectively. The achievement of this strength requirement is particularly more
critical at column intersections, which form the weakest areas.

3 CASE STUDIES

Two case studies of floating sheet pile cofferdam with different excavation depth are described. Exca-
vation depths of the respective case studies are 13 and 16 m. The observed field performances of the
two excavations and a comparison with the theoretical predictions are presented in the following sec-
tions. The analyses of the retaining systems were carried using PLAXIS, a finite element code for
geotechnical applications. The undrained shear strength (cu) for the jet grout layer is 600 kPa and E-
value is set as 100Cu, and 250 kPa as the tensile strength. The design undrained shear strength profile
of the marine clay is assumed to increase linearly with depth at a ratio of 0.16γ’z starting from 15 kPa.
The undrained modulus of the marine clay is based on an empirical correlation of 200cu. The design

328
undrained modulus of the marine clay is based on an empirical correlation of 200cu. The design con-
siders a surcharge of 20 kN/m2 with the ground water table located at 1.5 m typically.

3.1 Design consideration

3.1.1 Case 1

CASE 1 refers to a section of an excavation for branch tunnel at Site A. Layout of the excavation
showing the temporary ground support, internal bracing and instrumentation is shown in Figure 4.

The excavation depth for this section of the cut-and-cover tunnel is 13 m. Marine clay is present at
about 9 m below the reclaimed fill and it extends to depth of 30 m, below the ground surface. Thick-
ness of the soft marine clay below the final excavation level is at least 24 m. More than three lengths
of sheet piles to embed the piles into the medium dense layer, and the use of a very heavy combined
sheet pile and soldier pile sections with at least 2 jet grout layers would be required.

A floating combined sheet pile cum soldier pile cofferdam with a jet grout layer at final excavation
braced internally with 3 layers of strut was therefore adopted for the construction of the tunnels.
Length of the combined sheet pile/soldier pile wall is 15 m with a penetration depth of 3.5 m below the
final excavation. The design thickness of the jet grout layer is 2 m initially but was increased to 2.5 m.
The jet grout layer is restrained against upheaval force by two rows of steel H-piles.

The tension H-piles are spaced at about 6.0 m center-to-center in both the longitudinal and transverse
directions. The design penetration depth of the H-piles into the very dense or hard soil stratum with N-
value of 75 blows per 30 cm or higher is 7.5 m. Preboring to facilitate installation of the H-piles is
foreseen and the requirement is to backfill prebore with lean concrete up to the jet grout layer to ensure
proper transfer of the upheaval forces.

Fig. 4 Layout and section of the excavation at Site A (CASE 1)

329
3.1.2 Case 2

CASE 2 describes an excavation at Site B for a junction between a main tunnel and a branch tunnel.
Layout of the excavation showing the temporary ground support, internal bracing and instrumentation
is given in Figure 5.

Fig. 5 Layout and section of an excavation for a junction between a main tunnel and a branch tunnel at Site B
(CASE 2)

The excavation depth is 16 m generally and 18.5 m locally at the center section. Marine clay is present
at about 9 m below the reclaimed fill and it extends to depth of 30 m, below the ground surface. Thick-
ness of the soft marine clay below the final excavation level is at least 24 m.

The proposed temporary earth retaining system is also a floating combined sheet pile cum soldier pile
cofferdam with a jet grout layer at final excavation, similar to Case 1. The cofferdam is braced inter-
nally with 5 strut layers generally and 6 layers locally at the deeper section. The designed length of the
combined sheet pile/soldier pile wall is 20 m with a penetration depth of 4 m below the final excava-
tion at the 16m deep section and 24 m length with 6 m penetration at the 18.5 m deep section. The 3.0
m thick jet grout layer is restrained against upheaval force by two rows of steel H-piles. The tension
H-piles are spaced at about 4.5 m center-to-center in the transverse direction and generally at about 3.7
m in the longitudinal direction. The designed penetration depth of the H-piles into the very dense or
hard soil stratum with N-value of 75 blows per 30 cm or higher is 7.5 m.

4 DISCUSSION

The observed behaviors of the two cases of braced excavation are described. The instrumentation for
monitoring the actual performance of the excavation at this site is rather limited. Inclinometers, piezo-
meters and settlement markers were the only provisional items specified in the tender document while
extensometers and strain gages, which are also essential instrumentation for monitoring deep excava-
tion, were not included.

Regardless, although the amount of instrumentation may be limited, but given the rather unconven-
tional support method and scarcity of such case histories, a descriptive account of the observed behav-
ior of braced excavations would still be of benefit. The key indicators used in comparing the actual be-
havior of the excavation with design are lateral ground movement, basal heave, visual observation of

330
the structural performance of the wall, walers and struts, and the jet grout layer where possible. The
calculated lateral ground movement or wall deflection is obtained from the numerical results using
PLAXIS.

Summary of the strengths of the jet grout piles is given in Table 1. Core samples were taken from the
installed jet grout columns and tested at 28 days. The shear strength of the jet grout column is deter-
mined from compression strength of the unconfined compression test and tensile strength is interpreted
from an in-direct method using Brazilian Test.

Table 1 Summary of Shear and Tensile Strength of Jet Grout Core Samples at 28 days
Location Shear Strength (kPa) Tensile Strength (kPa)
Case 1 - B1-1: Zone B2 (2 samples) 789 426
Case 2 - M2 (7 samples) 683 257
Average value based on 15 samples 774 339

4.1 Actual field performance

4.1.1 Case 1

The as-built conditions of the cofferdam at Site A differ from the design in various aspects. The key
differences are: firstly, thickness of the jet grout layer was 2 m instead of the required 2.5 m, secondly,
the H-piles were installed by driving to set only, and lastly, the removal of the top 1.5 m by open cut
was done only on one side of the excavation. The actual penetrations of the H-piles vary between 32
and 35.5 m (i.e., between RL 71.0 m and RL 67.5 m). In relation to the prevailing ground condition,
the H-piles were terminated in the fluvial deposit.

At a local area near to the existing MRT tunnels, jet grouting was carried out in a controlled manner,
and with the use of casing and precutting, impact on the MRT tunnels was effectively kept to a mini-
mum. Data from automatic total station showed lateral movement of the MRT tunnels of up to ±5 mm
during the whole jet grouting operation. The measured data is given in Figure 7.

Schedule of the excavation sequence is given in Fig. 6. Excavation started on 30 November 2001, ap-
proximately one month after completion of the jet grout layer. Staged excavation and installation of
the strut layers progressed fairly smoothly to the 3rd excavation level according to the design. Summa-
ries of measured lateral earth movement adjacent to the excavation at inclinometers B1-I13 to B1-I16
are plotted below the excavation schedule in Fig 6. Profiles of the calculated wall deflection, shown as
dashed line. The comparison showed poor agreement between the calculated and the measured both in
term of the magnitude and the deformed shape. The FEM generally predicts larger lateral movement.
A smaller difference is however noted at B1-I13. The inclinometers showed a trend of increasing lat-
eral ground movement below the jet grout layer starting from 2nd excavation.

Although extensometers B1-E07 to B1-E09 were installed, the collected data showed almost no
change in movement, quite contrary to expected behavior. The instrument insensitivity could be due
installation problem or wrong positioning of the instrument, which were all located next to the H-piles
to avoid being damage during excavation. Hence, data from the extensometers are not presented.

Work was temporarily stopped over the Lunar New Year of 2002. On 16 February and about one
month after the completion of the 3rd strut layer, excavation for the 4th layer resumed starting in the
area furthest away from the MRT tunnels. Site workers reported several strange explosive-like sounds
in the early morning of 17 February. On the morning of 17 February, the condition of the cofferdam is
as shown in Photograph 1. Photograph 2 showed a close-up view of the condition of the collapse sec-
tion.

Photograph 3 showed a stockpile of excavated earth spoil adjacent to the collapsed section of the ex-
cavation. Visually, the observed height of the stockpile is likely to be more than 6 times the equivalent
height of the 20 kPa surcharge allowed for in the design.

331
 Ex c av at io n Seq uenc e
103
Ex c av at io n

100 Ins t all s t r ut

1

97
2

94
3
4
91

D at e

Inclinometer B1-I13 at B1-1 ZoneB Inclinometer B1-I14 at B1-1 Zone B Inclinometer B1-I15 at B1-1 Zone B Inclinometer B1-I16 at B1-1 Zone B
Lateral Movement toward Excavation [mm] Lateral Movement toward Excavation [mm] Lateral Movement toward Excavation [mm] Lateral Movement toward Excavation [mm]

-60 -40 -20 0 20 40 60 80 -10 0 10 20 30 40 50 60 70 -10 0 10 20 30 40 -20 -10 0 10 20 30 40

0 0 0 0

5 5 5 5

10 10 10 10

15 15 15 15

20 20 20 20

25 25 25 25

30 30 30 30

35 35 35 35

Initial Initial
40 1st layer 40 1st layer 40 40
2nd layer 2nd layer Initial Initial
3rd layer 3rd layer 1st layer 1st layer
1-D after collapse 1-wk before collapse 2nd layer 2nd layer
45 45 45 45
1st - cal 1st-cal 3rd layer 3rd layer
2nd - cal 2nd - cal 1-D after collapse 1-D after collapse
3rd - cal 3rd - cal 8-D after collapse 8-D after collapse
50 50 50 50

Fig. 6 Excavation schedule and summaries of inclinometer data B1-I13 to B1-I16 at Site A (Case 1)

Photograph 1 Collapse of a section of the sheet pile wall

332
 Photograph 2 A close-up view of the collapse section showing the ruptured jet grout layer and conditions
of the H-piles, struttings and the sheet pile wall opposite the collapsed section

Photograph 3 Condition of the stockpile of excavated spoil adjacent to the collapsed section

Immediately after the collapse, no further data was reported for inclinometers B1-I13 and B1-I14. At
inclinometers B1-I15 and B1-I16, the change in lateral movement at 1 day after the collapse is practi-
cally negligible and at 8 days after the collapse, an increase of less than 5 mm was reported. The in-
crease in movement is insignificant in comparison to that at the collapse section. Impact of the collapse
of the retaining wall on the existing MRT tunnels as shown in Figure 7, is also quite insignificant. The
change in lateral movement of the tunnels is not more than 2 mm.

The conditions of the collapse could be summarized as follows:

• The failure happened in a sudden and abrupt manner.

• The collapse occurred close to the corner of the cofferdam and primarily on one side (the side
adjacent to the high stockpile of excavated earth spoil).
• The jet grout layer ruptured distinctly at mid-span of the cofferdam.
• The H-piles were displaced upward by more than 3m above the existing ground level.

333
 Je t Gr o u t In s tallatio n Excavatio n Stag e A ft e r C o llap s e o f Sh e e t p ile s

-6 -6 -6

-5 -5 -5

-4 -4 -4

-3 -3 -3
Lateral dis plac ement [mm]

Lateral dis plac ement [mm]

Lateral dis plac ement [mm]

-2 -2 -2

-1 -1 -1

0 0 0

1 1 1

2 2 2

3 3 1 s t la ye r - m ax 3
1 s t la ye r - m in
A t c o lla ps e - m a x
4 4 2 nd la ye r - m ax 4
2 nd la ye r - m in A t c o lla ps e - m in
J G - m ax
5 5 3 rd la ye r - m ax 5 L o n g t erm - m a x
J G - m in
3 rd la ye r - m in L o n g t erm - m in
6 6 6
59680 59630 59580 59530 59480 59680 59630 59580 59530 59480 59680 59630 59580 59530 59480
Chainage Chainage Chainage

Fig. 7 Impact of the jet grouting operation and the collapse of the retaining wall at Site A on existing MRT
southbound tunnel

• The rupture of jet grout layer occurred right at the center of excavation (i.e. at mid-span section).
• The mode of movement of the collapse section of the retaining wall was rotational with the toe
swinging inward towards the excavation.
• The 3rd, 2nd and 1st strut layers had either buckled and/or displaced diagonally upwards.
• Buckling of the strut in the vertical direction could be due to the pull-out force from the H-piles.
• The movement of the opposite wall, across from the collapse wall and adjacent to the existing
MRT tunnels, if any is significantly small in comparison (see Figure 8).

A schematic of the collapsed condition of the cofferdam is shown in Figure 8. The cause of failure is
usually a combination of a number of factors. One of the key factors leading to the failure could be at-
tributed to the high stockpile of earth spoil placed too close to the excavation. However, the stockpile
has been placed there for sometime. The failure could be triggered by the commencement of the exca-
vation of the 4th layer. The several explosive-like sounds heard by the workers could be related to the
flexural failure of the jet grout layer, the buckling of the 3rd layer and subsequent strut layers, and the
pull-out of the H-piles. The fact that the H-piles were pulled out from the excavation indicates the H-
piles did provide a positive restraint to the jet grout layer but could not sustain the load when its pull
out resistance was exceeded. With the loss of restraint by the H-piles, the full earth load was then
transferred onto the full span of the jet grout layer causing it to buckle at the mid-span accompanied by
the circular slip failure of the adjacent ground and inward rotation of the sheet pile wall.

Sheet P iles
Sheet P iles
R eclaim ed Sand Fill

V ery Soft to Soft M arine

C lay
JG L A Y E R

Fluvial Sand / C lay

M edium D ense to V ery

D ense to H ard O ld
A lluvium

Fig. 8 A schematic of the failure pattern of the temporary sheet pile earth retaining wall at Site A

334
4.1.2 Case 2

At Site B, sheet piles and soldier piles were already installed and installation of the jet grout piles was
still in progress at the time of the collapse incident at Site A. The collapse incident came in timely for
the implementation of proper installation of the tension H-piles by preboring to achieve the required
pile penetration into the very dense or hard layer.

In order to ensure proper transfer of the upheaval force onto the H-piles, some additional measures was
implemented. These are: addition of two pair of shear links at 150 mm spacing to the section of the H-
pile within the jet grout layer, and 2) jet grouting in two stages. In stage 1, majority of the jet grout
columns was installed except those at the H-pile location. In stage 2, the omitted columns around the
H-piles were installed deeper by 500mm after installation of the H-pile. The improved detail of the
tension H-pile at the interface with the jet grout layer is also given in Figure 4.

The cofferdam was generally constructed in accordance to the design requirements. The installed
sheetpile/soldier pile length was 24 m, 4 m longer than the designed length. Concerns on the Main
Contractor’s strut layout for the junction area with many re-entrant corners was highlighted and em-
phatically conveyed by Designer to the Contractor.

Schedule of the excavation sequence is shown in Figure 9. Excavation started on 11 May 2002, at least
one month after completion of the jet grout layer. Staged excavation and installation of the strut layers
progressed fairly smoothly to up to the 3rd strut layer. It was at this time that the omission of the instal-
lation of monitoring instrumentation was realized. Work was then temporarily suspended until the re-
quired instrumentation including inclinometers, extensometers and settlement markers were installed.
Excavation resumed and continued down to the final excavation.

Summaries of the measured lateral ground movement and basal heave are presented in Figures 9 and
10, respectively. Profiles of the calculated lateral ground movement adjacent to the wall at various ex-
cavation stages are also plotted as dashed lines in Figure 9. Inclinometer data were only recorded after
completion of the 2nd strut layer. The comparison in general showed poor data match between the cal-
culated and the measured. The actual measured movements at various excavation stages were not only
significantly smaller except that at the final excavation, there is also a marked difference in the de-
formed shape. With some adjustment to the inclinometer data MI-06, the measured data on 28 August
yields a better agreement with that calculated in terms of magnitude and the deformed shape. As for
the extensometers, measurements were only taken after the completion of the 3rd strut layer. The re-
sults given in Figure 10, showed a trend of continual increase at ME-08 and quite an abrupt increase
from 4th to 5th excavation at ME-07 but tailed off thereafter. The calculated heave is much smaller than
the measured. The calculated total heave seems to give better agreement if the missing initial heave is
included.

From Figure 9, the measured data on 28 August, approximately two months after final excavation at
the 16m deep section and one month at the 18.5 m deep section, showed significant increase in lateral
ground movement below the jet grout layer, particularly for MI-07 though the excavation depth is
shallower. With missing data from the 6th to final excavation, it is difficult to know the sudden in-
crease is triggered at what excavation depth. Clearly, the trend of both inclinometers seems to suggest
that significant inward movement of the adjacent ground towards the excavation could be triggered at
a critical excavation depth. The usually high lateral movement at MI-07 reported on 28 August could
also be due to other factors such as consolidation or localized excavation for pile cap.

The potential for basal heave failure has not posed to be a problem at this excavation though larger
than expected incremental increases in basal heave were noted. The measured increases in basal heave
at the various excavation stages, as can be seen in Figure 10, are generally higher than that predicted
from FEM. In order to limit further increases in basal heave and to avoid another case of possible fail-
ure, a change in construction sequence was deemed to be prudent. The original intent is to complete all
the excavation and then built-up the permanent structure from the deepest section upwards. The re-
vised sequence involve the completion of the base slab at the shallower section on each end first and
then proceed with further excavation in the center which is also the deepest section.

335
 E x c av at i o n S eq uenc e
103
E x c av at i o n
100
2 Ins t al l s t r ut
97
94 3

91 4

88 5
6
16m
85
1 8 .5 m
82

D at e

Inclinometer M -I07 Inclinometer M -I06 M easured vs Calculated - 16m Excav M easured vs Calculated - 18.5m Excav
Lateral M ovement toward Excavation [mm] Lateral M ovement toward Excavation [mm] Lateral M ovement toward Excavation [mm] Lateral M ovement toward Excavation [mm]

-10 0 10 20 30 40 50 60 70 -20 0 20 40 60 80 100 -20 0 20 40 60 80 100 -20 0 20 40 60 80 100

0 0 0 0
28-Aug-02 28-08-02:mod
Final-cal Final-cal

5 5 5 5

10 10 10 10

15 15 15 15

20 20 20 20

25 25 25 25

30 3rd layer 30 30 30
4th layer 3rd layer
5th layer 4th layer
28-Aug-02 5th layer
35 3rd-cal 35 28-Aug-02 35 35
5th-cal 3rd-cal
4th-cal 4th-cal
Final-cal 5th-cal
40 40 40 40

Fig. 9 Excavation schedule and summary of inclinometer data MI-06 and MI-07 at Site B (Case 2)

ME06 ME08 ME07

100 100 100
20.741 18.396 18.281 95
22.775 24.066 20.391
88 22.451
24.802 24.826
80 26.795 80 Cal, wrt 3rd 80 24.231 82
Cal, wrt 3rd 76
Vertic al Mov em ent [m m ]

Vertic al Mov em ent [m m ]

Total @final 74
Vertic al Mov em ent [ m m ]

Cal, wrt 3rd

Total @final 72
60 68

60 60 62
60
50 59

55 56
50
49

45
40 40 40 40
37

35
31
29 28
29

25 24
21 21
21

20 20 20
18
17
17

15 16 16
9 16

14 15 15
14

13
11
9
5
0 0 0 0 0
0

Date Date Date -5

05-06-02

12-06-02

19-06-02

26-06-02

03-07-02

10-07-02

05-06-02

12-06-02

19-06-02

26-06-02

03-07-02

10-07-02

05-06-02

12-06-02

19-06-02

26-06-02

03-07-02

10-07-02

Fig. 10 Summary of extensometer data ME-05 to ME-08 at Site B (Case 2)

336
5 CONCLUSIONS

Two case studies of floating-type braced excavation in soft marine clay were described. The two cases
with different excavation depth and at different location within the same site have fairly similar ground
condition consisting of reclaimed fill over very thick layer of marine clay.

The first case study with 13 m excavation collapsed abruptly when excavation for the 4th strut layer re-
sumed just after the festive break for the Lunar New Year and about one month later after completion
of the 3rd strut layer. A number of factors could have contributed to the failure.

• Stockpiling of earth spoil to height of at least 6 times the equivalent height of 20 kPa surcharge
placed too close to the collapsed section of the excavation,
• H-piles were installed by driving to set only instead of the designed penetration of 7.5 m into the
very dense or hard layer.
• From Photographs 1 and 2, it is also evident that only one row of tension H-piles while the design
requires two rows of H-piles.
• Although the H–piles were only driven to set, the extent of the pull-out of the H-piles would sug-
gest that the piles did provide a positive restraint to the jet grout layer.
• With the loss of the restraint provided by the H-piles, the full span of the jet grout layer which
could only sustain a marginal load then failed by buckling or shearing right at the mid-span when
the full earth load is transferred onto the jet grout layer.
• The failure of the jet grout layer leads to the development of full slip circle failure of the adjacent
ground accompanied by inward rotation of the sheet pile wall, and buckling and displacement of
the strut layers.

The second case study with 16 m excavation and 18.5 m deep locally was located in ground with a
thicker and deeper marine clay layer was completed successfully. At this excavation, the tension H-
piles were installed 7.5 m into the very dense or hard layer by preboring and backfilled with lean con-
crete in the lower section and earth in the upper section. Jet grouting was carried out in two stages to
ensure proper force transfer from the jet grout to the H-piles. Significantly higher increase in basal
heave was reported. An increase in the lateral movement below the jet grout layer was also noted. The
increase is small and gradual but an abrupt increase was reported sometime after reaching the final ex-
cavation, particularly at M-I07. Although measured ground movements both lateral and base heave,
were greater than expected, there was no indication of any distress to the exposed jet grout layer. The
construction sequence was revised to limit further increase in ground movement and the permanent
structure has since been completed.

One key lesson learnt from these two case studies is the recognition of the important role of the tension
H-piles in reducing the flexural stresses of the jet grout layer. The other is not to stockpile excessive
earth spoil too high and too close to the excavation. The adapt: “penny wise pound foolish” derived
from considerable wealth of past experiences is indeed a proven wisdom.

ACKNOWLEDGEMENT

The authors wish to thank Mr. Mya Sein for his assistance in gathering all the necessary information
pertaining to the project.

REFERENCES

Mana, A.I. & Clough, W.G. 1981. Prediction of movements for braced cuts in clay. Journal of the Geotechnical
Division, Proceedings of the American Society of Civil Engineers, Vol. 107, No. GT6, 759-777.
Weissenbach, A. & Kempfert, H.-G. 1995. German national report on ‘braced excavation in soft ground’. Un-
derground Construction in Soft Ground, Fujita & Kusakabe eds, 33-36, Rotterdam: Balkema.
Guilloux, A. & Kastner, R. 1995. German national report on ‘braced excavation in soft ground’. Under-ground
Construction in Soft Ground, Fujita & Kusakabe eds, 29-32, Rotterdam: Balkema.

337