Design and Construction of A Three-Sided ERSS Wall For Launch Shaft
Design and Construction of A Three-Sided ERSS Wall For Launch Shaft
Design and Construction of A Three-Sided ERSS Wall For Launch Shaft
ABSTRACT: Thomson East Coast Line Contract T202 comprised of excavation works to construct
an overrun tunnel, station, crossover tunnel and a TBM launch shaft (LS1). Open cut slopes were
adopted for all areas except LS1, which is adjacent to the crossover tunnel excavation at the southern
direction. The size of LS1 is 30m(W) x 30m(L) x 16m(D), where the original proposed Earth Retain-
ing Structural System (ERSS) was to adopt a conventional four sided ERSS walls with multiple steel
struts. This original scheme was time consuming to the contractor as concurrent excavation for both
LS1 and crossover tunnel is not allowed at the same time. An excavation support system with only
three sides of structural ERSS walls and one open end interfaced with open cut slopes was eventually
adopted. In doing so, the design of the LS1 would need to cater for imbalance forces from the ERSS
walls and the load has to be transferred effectively so that the ERSS walls will have sufficient re-
sistance to reach equilibrium. This paper discusses the design challenges and consideration in provid-
ing a safe and robust ERSS system for LS1 excavation for T202 and also discusses instrumentation
data measured on site.
1 INTRODUCTION
The most typical construction method for launch shafts is to construct a rectangular shape braced cof-
ferdam, which is retained by Earth Retaining and Stabilizing System (ERSS) Walls at all sides. Intui-
tively, this would be the choice as there are sufficient past experiences in design and construction,
which is an important factor to ensure a safe and robust ERSS system. However, when there is an in-
terface with open cut slopes excavation immediately adjacent to a braced excavation, a rectangular
cofferdam closed at four sides may not be the best option to be pursued, as concurrent excavations
may not made possible due to incompatibility in the interface between open cut slopes and braced cof-
ferdams.
This paper will discuss the experience faced when it comes to such condition, where open cut slopes
were adopted next to a launch shaft using braced cofferdams in Contract T202 in the Thomson East
Coast Lines Contract. This paper will highlight the factors that leads to the selection of the ERSS op-
tions, and the design considerations being made to ensure that the ERSS system is efficient and ro-
bust.
The Thomson East Coast Line Contract T202 covers the construction and completion of Woodlands-
North-Station (WDN), at grade overrun, cut and cover overrun tunnels, cut and cover crossover tun-
nels south at station, a TBM launch shaft (LS1) and two mainline bored tunnels from launch shaft
LS1 to Woodlands Station (T203).
The WDN is located on a hilly site, as shown in Figure 1. The main structure is underground, and
hence the ERSS works formed an important part of the project. Since there are ample space to form
the slopes within the site boundary, the general ERSS scheme adopted at this site is open cut slopes.
The deepest depth of excavation, which is located at the east side of the WDN station, is about 42m
from the existing level.
Launch Shaft (LS1) is located at the southern end of the project, from which two tunnel boring ma-
chines (TBM) of 6.6m diameter were later driven towards T203 station towards the south direction.
Figure 1. Aerial View of Contract T202 Construction Site, Launch Shaft at Far Right
In general, the site consists of ~3 to 7m layer of fill, followed by Bukit Timah Formation residual soils
with increasing SPT N value. Typically, there are no GIV layers, and the GVI or GV residual soils
will followed by a solid GIII layer. The soil design parameters from the GIBR are shown in Table 1.
The original ground level at LS1 is about RL130m. Due to sufficient space, the contractor decided to
cut down the original ground level to RL118m – where the capping beam of LS1 will be started. This
would also reduce the total length of the ERSS walls, and also to reduce the number of struts. Full
open cut slopes down to the final excavation level of RL102m within the launch shaft is not possible,
as the existing Republic Polytechnic Carpark is located at the South-East Direction of LS1.
As such, after excavate down to RL118m, the Fill layer will be excavated away, which subsequently
leave the soil behind the ERSS walls to be at least G(VI) material and above.
2.3 Launch Shaft Original Design
The original design with a more or less squared shape contiguous bored pile (CBP) cofferdam, as
shown in Figure 2, was proposed. The interpolated rock head profile based on the tender stage soil in-
vestigation information projected onto the CBP wall elevation is shown in Figure 3. At the north-west
of the LS, the rock levels are high enough that structural ERSS walls were not required and rock
slopes were proposed to act as ERSS walls. The TBMs will then be launched and break-out through
the Southern Wall.
Figure 2. Original Intended Launch Shaft ERSS Layout Plan, with horizontal struts strutting across North Wall
and South Wall
Figure 3. Inferred Elevation View of Rock Profile during Initial Stage ERSS scheme development,
with 4 sides of Wall
The rock-head contours are shown in Figure 4. The difference of rock-head levels of GIII or better
was found to be in the range of 30m, from RL118 to RL88, just across a short horizontal distance
within the LS excavation, across the four (4) sides of the ERSS wall.
Figure 4. Rock-Head Contour within the Launch Shaft
Figure 5 shows the layout plan of the original ERSS wall layout for launch shaft in relative to the
Crossover Tunnel Excavation, which is using open cut slope method.
Should the original design to be adopted, the intended construction sequence is to excavate the LS1
first. This sequence requires that the soil/rock behind the North Wall, which act to counter the lateral
force from the South Wall through struts as shown in Figure 6, to only be removed after the comple-
tion of the LS1. In other words, the construction of the structures at the interface between Crossover
box and LS1 will be left last. This would affect the overall basic structure completion program. This
problem is amplified as the berm consists of GIII or better rocks, which are very difficult to excavate
and remove.
Launch Shaft
LS1
East Wall
South Wall
North Wall
West Wall
2.5 Proposed Three (3) Sided ERSS wall with Open-Ended Excavation
The sequence of the ERSS in its entirety and its interface with the Crossover box ERSS were re-
evaluated. Several proposals were discussed, including shifting the position of the Launch shaft to a
location were change in rock head profile is not so drastic, or the use of circular shape cofferdam.
Eventually, a C-shaped cofferdam with open cut excavation at the north end of the launch shaft was
adopted. The plan and sectional view of the LS1 adopting the revised ERSS Scheme are shown in
Figure 7 and Figure 8, respectively. With this scheme, the excavation of the Launch shaft and crosso-
ver can be carried out at the same time. The elevation view of the wall is shown in Figure 9, where the
original North Side wall can be seen omitted.
West Wall
North Wall
(Omitted) South Wall
East Wall
Figure 7. Launch Shaft for T202, with open end at the North Direction
South Wall
Figure 8. Elevation View of the South Wall in the revised ERSS Scheme (Open Ended Construction)
First, the earth pressure into CBP Wall will be transferred into the Diagonal Struts. The Diagonal
struts will than transfer the load into waler members. To ensure that forces can transfer efficiently,
shear connections are provided between diagonal struts and walers.
Waler members than will transfer the load into the CBP wall at the East Side. Finally, the CBP has to
transfer the load back into the ground. To achieve equilibrium, the CBP wall as to be ensured that
there are sufficient shaft resistance between soil/rock and CBP wall.
A simple and crude way to estimate the imbalance forces is from the predicted strut forces for every
level obtained from the geotechnical analysis. Then, a total value from all the strut loads will be
summed up. Table 2 shows the summation of strut forces for East Wall. For West Wall, the same con-
cept is used. The strut forces shown here is in per meter run, where this value will be multiplied by
half of the width of the excavation, assuming that the half side of the earth forces from South Wall
will transfer to East Wall, and the other half to West Wall.
Table 2. Predicted forces transferred to East Wall based on Individual Stages (From Drained Case)
_________________________________________________________________________________________________________
State S1 S2 S3 S4 Total force Width Max forces
at individual considered transferred to
stage (kN/m) (m) East Wall (kN)
_________________________________________________________________________________________________________
Exc to 116.0 -- -- -- -- -- 17.5 0
Exc to 112.0 432 -- -- -- 432 17.5 7,560
Exc to 109.5 254 683 -- -- 937 17.5 16,398
Exc to 104.0 158 366 1186 -- 1710 17.5 29,925
Exc to F.E.L. 180 421 1312 543 2456 17.5 42,980
Remove S4 125 307 1370 -- 1802 17.5 31,535
Remove S3 142 1022 -- -- 1164 17.5 20,370
Remove S2 490 -- -- -- 490 17.5 8,575
__________________________________________________________________________________________________________
3.3 Geotechnical Resistance to cater for Imbalance Forces
The geotechnical resistance provided by shaft friction between soil/rock and CBP wall is based on the
contact area between the CBP wall and the surrounding soil. To simplify calculation, only a planar
contact area is considered. Therefore for each pile the area which is considered as follow:
Qs = Diameter of CBP wall x Height of Soil/Rock in contact with pile x shaft friction
= 1.5m x H x fs
Due to undulating nature of rock-head surface, the piles are divided into sub-groups for simplification
purposes, as shown in Figure 11.
The total resistance from soil and rock is shown in Table 3, summed up to be 96,708kN. The factor of
safety would be 96,708kN/ 42980kN = 2.25, which is satisfactory.
Table 3. Total Geotechnical Resistance from Soil and Rock at the Eastern Wall side
_________________________________________________________________________
Pile Soil /Rock Resistance
_________________________________________________________________________
1 to 5 Soil 900
6 to 13 Soil 2,160
14 to 19 Soil 3,060
20 to 21 Soil 1,320
22 to 29 Soil 10,608
1 to 29 Rock (frm Rock Socket) 78,300
________________________________________________________
TOTAL 96,708
From this simple limit equilibrium calculation practice, the proposed ERSS Wall with 3-sided CBP
and one open-end excavation is feasible.
The geological formation in T202 is Bukit Timah Granite Formation, which is well known for its var-
iability in the bedrock level and attributed to the deep tropical weathering explained by Shirlaw et al
(2000). One key risk associated with excavation in the Bukit Timah Granite is through its fractured
rocks where the ERSS walls may not sufficiently ensure adequate hydraulic cut-off. In common exca-
vations, groundwater drawdown may cause potential consolidation settlement should there is soft soil
material. Therefore it is common that mitigation measures to minimize this risks are adopted in areas
where building settlements are a concern.
However, the conditions are different from common excavation projects, where excavation of open
cut slopes adjacent to LS1 could cause water drawdown, which will cause water loss at the crossover
box direction.
In view of this, it is therefore decided not to provide any hydraulic cut-off to the CBP walls itself. The
decision in this is due to the following reasons:
a) Since there is considerable groundwater loss through adjacent open cut slopes, providing hy-
draulic cut-off in the CBP walls may not be useful.
b) As open cut slopes is adopted in the adjacent area, promoting groundwater drawdown should
be the correct way in order to provide a safe open cut slope.
It is important to note the impact of the water drawdown due to the open cut slopes towards any build-
ing structures has been ascertained, and there is no adverse impact towards the nearest building struc-
ture, which is considerably a distance away of 60m from edge of excavation.
Altogether, there are five (5) nos of In-wall Inclinometers and three (3) nos of In-soil inclinometers.
38 numbers of settlement markers, five (5) water standpipes and five (5) piezometers.
The excavation of LS-1 started on April 2015 and reached the final excavation level on August 2015.
Upon casting the base slab, the 4th level struts were removed on August 2015 in order to carry out the
Tunneling works. The 1st level to 3rd level struts were then remained for a period of more than two (2)
years, which the 3rd level struts only started being removed on November 2017. All the struts were
completely removed on March 2018. Generally, inclinometer readings in the A-Direction (towards
excavation) shows similar trend to FEM prediction, but with a lesser magnitude, as shown in Figure
13 for In-wall inclinometer IW2403.
Figure 13: Comparison of inclinometer readings with FEM prediction in A-direction for East Wall
The magnitude of the ERSS wall movement in both direction are shown in Figure 14, at the stage of
removal of all struts. It is observed that the East wall has larger movement compared to the West
Wall. One particular reason for this is at the West Side, the ground material comprised of GIII rock,
while at the East Side the ground material is residual soil.
It is also interesting to note that the ratio of the wall deflection in B-direction to the A-direction is
greater than 0.5, as briefly summarized in Table 4. This shows that the movement in B-direction is
significant and should not be taken lightly during the monitoring stage. This also indicates that there is
an imbalance force which leans towards the open-end direction, which needs to be appropriately taken
into account during the design stages.
4 CONCLUSION
A three sided (3) ERSS wall with one open end was chosen to construct the Launch Shaft LS1 in
Woodlands North Station Construction. Conventional four (4) sides closed cofferdam was not adopted
as this method may prolong the overall construction sequence.
As there is an open end within the Launch Shaft, careful consideration need to be made during the de-
sign process to cater with the imbalance forces, especially the adequacy in geotechnical resistance
from the both sides of the open ended ERSS walls. A simple check of the load path and geotechnical
resistance to resist the ERSS walls were carried out and the scheme were found to be feasible.
The groundwater consideration in the CBP walls were slightly different from what is commonly prac-
ticed, whereby water drawdown is allowed without hydraulic cut-off. This is made possible due to the
consideration that there is an open cut slope just adjacent to the LS.
Instrumentation and monitoring program shows that there are lateral movements in B-direction to-
wards the open-ended excavation direction, which shows that there are imbalance forces towards this
direction, as expected. The LS excavation has been successfully completed, suggesting that the design
considerations made were align with the actual condition.
REFERENCES
K.H. Goh, T.F. Lim, D.C. Chen & D Wen. Excavation and Temporary Retaining Wall Systems for the
Stations of Downtown Line, Proc. Conf. Underground Singapore 2014
Shirlaw, J.N., Hencher, S.R. and Zhao J. 2000. Design and Construction Issues for Excavation and
Tunnelling in Some Tropically Weathered Rocks and Soils, Proceedings GeoEng2000, Melbourne,
Australia, 19-24 November 2000.