Conference PaperPDF Available

The choice of EPB or slurry shields for tunneling in mixed face conditions resulting from tropical weathering

  • WSP-Golder

Abstract and Figures

There has been a general trend in Singapore, over the last 30 years, for slurry shields to be increasingly used in preference to EPB, particularly when tunnelling in mixed faces of weathered granite. Initially this was because of problems with maintaining the required face pressure with EPB TBMs in mixed face conditions. With modern EPBs and proper conditioning, pressure can be maintained. However, EPB tunnelling can still be problematic, particularly where there is a high percentage of strong rock in the face. Two tunnel drives in mixed ground from the weathering of strong igneous rock, one by slurry TBM and one by EPB, are compared. This is done by reviewing the TBM parameters, in terms of Penetration Index and Specific Energy. The effect of the ground conditions on the instantaneous advance rates, disc consumption and time required for interventions is assessed. The nature of the problems specific to the EPB tunnelling are identified, and compared with documented examples from Singapore and elsewhere.
Content may be subject to copyright.
Pressurised Tunnel Boring Machines (PTBM) are used to construct tunnels in ground where a face
pressure needs to be applied to ensure stability and to control the ground movements due to tunnelling.
There are two basic types of PTBM, the slurry shield and the earth pressure balance (EPB) shield. In
the slurry machine (including the Mixshield type) the face pressure is applied through the medium of
a liquid slurry, commonly a bentonite slurry, which also acts as the medium for transport of the cut
ground, or spoil. In the EPB shield the pressure is applied through the medium of the cut ground
(spoil), with the properties of the spoil modified by the addition of conditioning agents.
In Singapore, there was extensive use of slurry machines for small diameter pipe-jacking from 1982,
when the first slurry TBM drive in Singapore was used to construct a 1500mm pipeline (Balasubra-
manian, 1987). However, medium diameter slurry TBMs were not used in Singapore until the con-
struction of parts of Circle Line 3 and 4, in 2005/6 (Nakano et al 2007).
The first Earth Pressure Balance Machine (EPB) used in Singapore was a 3.754m diameter TBM, em-
ployed to drive a 3 km long outfall pipeline from the Ulu Pandan Sewage works, between 1983 and
1985 (Balasubramanian, 1987). This was followed by two 5.910m diameter EPB TBMs used to drive
the Mass Rapid Transit (MRT) tunnels between Lavender and Bugis Stations, in 1986-1987 (Elias and
Mizuno, 1987). Subsequently, 14 EPB TBMs were used to drive the tunnels for the North East line
(1999-2001), and 8 EPBs for the Deep Tunnel Sewer System (DTSS), between 2001 and 2005 (Mar-
shall and Flanagan, 2007). Apart from the Ulu Pandan sewer TBM, all of these TBMs were mid-sized
(4m to 9m diameter).
No large diameter (>9m diameter) PTBM has yet been used in Singapore.
The choice of EPB or slurry shields for tunnelling in mixed
face conditions resulting from tropical weathering
J.N. Shirlaw
Golder Associates (HK) Ltd, Hong Kong
ABSTRACT: There has been a general trend in Singapore, over the last 30 years, for slurry shields to
be increasingly used in preference to EPB, particularly when tunnelling in mixed faces of weathered
granite. Initially this was because of problems with maintaining the required face pressure with EPB
TBMs in mixed face conditions. With modern EPBs and proper conditioning, pressure can be main-
tained. However, EPB tunnelling can still be problematic, particularly where there is a high percent-
age of strong rock in the face. Two tunnel drives in mixed ground from the weathering of strong igne-
ous rock, one by slurry TBM and one by EPB, are compared. This is done by reviewing the TBM
parameters, in terms of Penetration Index and Specific Energy. The effect of the ground conditions on
the instantaneous advance rates, disc consumption and time required for interventions is assessed. The
nature of the problems specific to the EPB tunnelling are identified, and compared with documented
examples from Singapore and elsewhere.
Underground Singapore 2016
Following the completion of the North East line and the DTSS tunnels, numerous projects for MRT,
cable and water tunnels, involving construction by medium-sized PTBMs, have been or are being con-
structed in Singapore. A summary of the medium diameter PTBMs used for completed and current
major projects in Singapore is provided in Table 1.
Table 1. A summary of the major projects involving medium sized PTBMs in Singapore
Major Projects
PTBMs: Numbers used
Tunnelling completed
East-West Line
North East Line (NEL)
Deep Sewer Tunnels
Circle line (CCL)
Downtown Line (DTL)
Deep Cable Tunnels
Under construction (2016)
Thomson-East Coast Line (TSL)
Under construction (2016)
Table 1 is not comprehensive, as a number of smaller projects, each requiring from one to five medi-
um diameter PTBMs, have been omitted.
A fairly recent development has been the production of mixed-mode PTBMs, which can be used in ei-
ther slurry or EPB mode. This type of TBM includes the variable density TBM. One mixed mode
EPB/slurry TBM was used in Singapore for the construction of a link sewer, connecting to the main
DTSS tunnel (Wallis 2005). This and the other TBMs used for the construction of the link sewers are
not included in Table 1, as these were all small (<4m diameter) machines.
Table 1 shows a major trend. This is a change, from all EPB TBMs before 2003, to increasing use of
slurry TBMs since 2003. However, this trend is not simply related to time, but also geology. The tun-
nelling has been mainly in five major geological strata:
1. The Singapore granite (grey in Figure 1)
2. The Jurong Formation, of meta-sedimentary rock (green in Figure 1)
3. The Fort Canning Boulder Bed, comprising quartzite boulders in a clayey-silt matrix (not shown
in Figure 1 as it does not outcrop)
4. The Old Alluvium (pink in Figure 1)
5. The Kallang Formation, including soft marine, estuarine and fluvial clay, and fluvial and beach
sands (brown in Figure 1)
The geological map of Singapore is shown in Figure 1, with the types of TBM typically used also
shown. As indicated in Figure 1, slurry shields have been used mainly for tunnelling in the Singapore
granite, with some use in the Jurong Formation. Medium diameter PTBM tunnelling in the soils of the
Fort Canning Boulder Bed, Old Alluvium and Kallang Formation has almost all been by EPB TBM.
Limited slurry tunnelling has been carried out in the Old Alluvium, but only where the selection of the
TBM is governed by encountering granite in other parts of the same drive.
The use of slurry shields for approximately 10.2 route km (20.4 km of tunnelling) of the Circle Line
was specified by the owner, the Land Transport Authority, and was mainly for tunnelling in the
weathered rock of the Singapore granite. The decision to use slurry shields in these conditions fol-
lowed experience of EPB tunnelling in the granite on the North East Line and the Deep Tunnel Sewer
System (DTSS). Particular problems had been encountered in mixed face conditions of soil and rock,
resulting from weathering of the granite.
For the North East Line, EPB pressure could not be maintained in the mixed face conditions in weath-
ered granite, as outlined in Shirlaw et al (2000). As a result there were a number of sinkholes and
ground losses.
EPB tunnelling in mixed faces of granite and saprolite for the DTSS was also problematic. Problems
For some of the mixed face tunnelling, the TBM could not be operated in conventional EPB
mode, but was instead operated in ‘semi-EPB’ mode, using compressed air above tunnel axis
The high compressed air pressure required for interventions, and the effect of the pressure on the
time required to change cutting tools. During the project, extensive use of dewatering was used
to reduce the compressed air pressure required for interventions
Frequent damage to cutting tools, and interventions to replace the tools
A new, and significantly redesigned, cutterhead for one TBM, installed during the tunnel drive
by sinking a special shaft
Major wear of a screw conveyor, requiring replacement during the tunnel drive (Figure 2)
A number of sinkholes (Shirlaw & Boone, 2005); an example is shown in Figure 3
Serious delays to the tunnelling schedule (Wallis, 2004); on the T05 drives, mainly in weathered
granite, actual production rates averaged about one third of those planned
Figure 2. Badly abraded screw conveyor, DTSS, sec-
tion in weathered granite
Figure 3. Sinkhole over DTSS EPB TBM in weathered
A further problem with EPB tunnelling in mixed ground conditions that has been noted on a number
of projects is that of high heat: the muck emerging from the screw conveyor has been measured at 60
to 80 degrees centigrade, locally, on some projects. The high heat has typically occurred when tunnel-
ling in a mixed face of soil and strong (or stronger) rock. As discussed in Shirlaw (2015) this high
heat can have a major, adverse, effect on the numerous interventions required in these ground condi-
tions. Tunnel staff cannot enter the excavation chamber due to the heat. In the examples cited it took
12 to 36 hours of flushing with bentonite slurry to reduce the heat to a level where it was safe to enter
the excavation chamber. The long period of flushing with the face stationary often results in erosion
of the face, and a partially caved face at the beginning of the intervention. The stability of the partial-
ly caved face deteriorates more rapidly during the intervention than would normally be the case.
The slurry TBM tunnelling for the Circle Line was not without its share of problems, as discussed
Merrie (2009). In particular, there was:
Local use of dewatering to assist the tunnelling
Frequent damage to cutting tools, and interventions to replace the tools
A major modification to the TBM, as discussed in Shirlaw and Hulme (2008)
Several sinkholes (see Figures 4 and 5 for examples); these were mainly associated with inter-
ventions of long duration and/or multiple interventions over a short length of tunnelling
Delay to the overall tunnelling schedule
Figure 4. Sinkhole over a slurry shield, Circle Line
Figure 5. Sinkhole over a slurry shield, Circle
Heat has not been a problem for slurry shield tunnelling in mixed ground conditions.
It is evident from Table 1 that despite the local problems with slurry tunnelling in mixed ground con-
ditions, there is still a strong preference for the use of slurry machines, rather than EPB TBMs, for
tunnelling in the granite.
Over the last 15 years there have been significant improvements in both EPB and slurry TBMs, and in
their operation, in relation to excavating in weathered rock. Merrie (2009) provides recommendations
for improving slurry TBMs for these conditions. The Land Transport Authority (LTA), in addition to
indicating the type of TBM to be used, provide specifications for EPB and slurry TBMs that set min-
imum standards.
However, mixed ground conditions comprising soil and rock remain challenging for pressurized TBM
tunnelling. In order to illustrate this, examples of EPB and slurry tunnelling in weathered rock are out-
lined, below. These two case studies are not from Singapore, but were in similar mixed ground condi-
tions, resulting from the topical weathering of strong (or stronger) rock, to those commonly encoun-
tered in the tunnelling in the Singapore granite. The tunnels were driven within the last five years,
using new TBMs, designed for the respective projects and the anticipated ground conditions.
Before assessing the interaction of the TBM and the ground, it is necessary to establish the actual
ground conditions encountered during tunnelling. In weathered rock, boreholes alone are a poor indi-
cator of rockhead, as shown in Fletcher (2004), Merrie (2009) and Shirlaw (2015). In the two case
studies of EPB and slurry tunnelling, the actual ground conditions were established from all of the in-
formation available, as discussed below.
Weathered bedrock profiles are highly erratic, even over short distances. Occasional, offline boreholes
can result in an inaccurate prediction of the actual rockhead profile (Shirlaw 2015). Above the rock-
head, corestone boulders have a major effect on tunnelling, but are notoriously hard to identify on the
basis of borehole data. However, once the tunnel is driven, a significant amount of additional data is
available to identify the ground conditions encountered. The information includes:
1. The original pre-construction boreholes and any additional boreholes drilled during tunnelling.
2. Face mapping; this can only be carried out during interventions into the excavation chamber
3. The TBM data, such as torque, thrust and rate of advance
4. Information from probing from the TBM (although this is rarely carried out during PTBM tunnel-
5. The nature of the excavated material
These are discussed further, below.
2.1 Borehole data
Boreholes provide a high degree of accuracy in terms of establishing the ground conditions at a par-
ticular location, particularly when double or triple tube core barrels are used to provide a continuous
core through the rock encountered. However, boreholes are commonly located some distance from the
tunnel alignment. Boreholes directly in the tunnel profile potentially create a path for slurry, com-
pressed air or foam to reach the surface, if not securely grouted. For this reason boreholes are typical-
ly located slightly off the tunnel alignment. However, in urban areas, the presence of buildings, utili-
ties and property boundaries often severely restricts where boreholes can be located, and they may end
up being located some distance from the tunnel alignment.
2.2 Face mapping
It is standard practice to prepare a record of the face conditions during an intervention. This is normal-
ly the only time during pressurized TBM tunnelling that the face can be inspected directly. However,
there are some limitations in the quality and nature of the data that can be obtained, including:
a) The face is visible only through the openings in the cutterhead. For a modern, mixed ground cutter-
head, these opening typically comprise only 25% to 35% of the area of the cutterhead
b) Most interventions are ‘half face’ interventions, with the muck in the chamber only drawn down to
axis level. If the rockhead is below this level, it will not be identified
c) It is common practice to form a filter cake on the exposed face, which may obscure the ground con-
ditions; however, the filter cake will not form on massive rock
d) Unless the tunnelling team includes a geologist or engineering geologist who is fit and trained for
work in compressed air, the quality of the mapping may be limited by the training of the personnel
preparing the face map
Despite these limitations, the face mapping records are generally the most direct and most accurate
records of the ground conditions encountered. However, these records can only be obtained at inter-
vention locations, and provide only local ‘snapshots’ of the geological conditions encountered.
2.3 TBM data
The data captured during driving the TBM provides continuous evidence of the geological conditions
along the tunnel drive. Numerous parameters are measured during TBM tunnelling. Three parameters
that are affected by ground conditions are: contact force on the cutterhead, the torque required to turn
the cutterhead and the advance rate, expressed either as mm /cutterhead revolution or as mm / minute.
Two functions that combine some of these parameters, and that have been found useful for assessing
tunnelling in weathered rock, are the Penetration Index and Specific Energy. The derivation of these
parameters is summarized below.
2.3.1 Penetration Index
The Penetration Index (PInd) can be calculated from:
PInd = CF/PRev
PRev is the penetration rate in metres per revolution
CF is the average normal force on a disc cutter in MN
2.3.2 Specific Energy
The Specific Energy (SE) can be calculated from:
Specific Energy (SE) =[(Fn.P) + (2.π.N.T)] / A.P MNm/m3 (or MJoules/m3)
Fn is the normal contact force on the cutterhead, in MN
P is the penetration rate in metres/minute
N is the rotation speed in revolutions per minute
T is the Torque in MNm
A is the cross sectional area of the TBM in m2
Although the Specific Energy includes terms that are based on contact force and torque, it is the latter
that is generally dominant.
2.4 Probing
Typically, medium and large sized PTBMs are equipped with ports that allow probing ahead of the
TBM. A full fan of peripheral grout ports typically allows drilling (and, if necessary, grouting) outside
the tunnel extrados; these ports are placed just behind the pressure bulkhead. There are also usually a
limited number of ports that allow drilling through the pressure bulkhead, directly ahead of the TBM;
probing can only be accomplished when an opening in the cutterhead is aligned with the port in the
In practice, tunnel managers are invariably resistant to probing from PTBMs; the author has never ex-
perienced probing through the bulkhead, and only very rarely from the peripheral ports. A major con-
cern with probing though the bulkhead is the risk of breaking the drill string, leaving steel ahead of the
TBM and, potentially, skewering the cutterhead.
2.5 The nature of the excavated material
The nature of the excavated material may provide some information on the ground conditions. The
value of the excavated material, as a basis for assessing the face conditions, is limited by a number of
factors, including:
Grinding and reworking of the material during cutting and then during the transport of the material
from the tool gap, into the excavation chamber, through the rock crusher (in slurry machines) and
then passage through the screw conveyor (EPB TBM) or along the discharge pipe and through the
slurry treatment plant (slurry TBMs).
Time delay between excavation and discharge at the surface
Addition of conditioning materials
These factors significantly limit the information on the geology of the tunnel drive that can be ob-
tained from the excavated materials. Pieces of rock on the conveyor belt (for an EPB machine) or sep-
arated on the shakers (for a slurry machine) show that the TBM has encountered some rock, but it is
generally not possible to tell the proportion of rock in the face with any degree of accuracy.
2.6 Overall summary
All of the sources of information listed above can provide information on the ground conditions that
the tunnel has encountered. In weathered rock conditions, it has been found that the quality of the data
can be ranked as follows, with the highest quality ranked as 1:
1. Face logs from interventions
2. Borehole data within 0.5 tunnel diameter from the tunnel extrados
3. TBM data, where calibrated against face logs and borehole data
4. Borehole data outside 0.5 tunnel diameter from the tunnel extrados
5. Data from probe holes and the excavated spoil
3.1 General
Two case studies are presented: Tunnel A (a slurry TBM drive) and Tunnel B (an EPB TBM drive).
Tunnel A was driven through weathered granite with intrusive dykes of basalt and rhyolite, while
Tunnel B was driven through weathered Tuff. Only a part of each tunnel drive is presented; these por-
tions have been selected to include saprolite (soil grades of weathered rock), mixed rock and saprolite
and full face tunnelling in rock. Neither of the tunnels was in Singapore. However, the case studies are
from conditions similar to those commonly found in Singapore, with deep tropical weathering, and are
presented here to compare the performance of slurry and EPB TBMs in these conditions.
The TBM information for Tunnels A and B is summarized in Table 2.
Table 2. The TBMs used to drive tunnels A and B
Tunnel A
Tunnel B
TBM Type
TBM cut diameter
Disc size
19” (482.6mm)
17” (431.8mm)
Number of discs
Opening ratio (%)
The Uniaxial Compressive Strength (UCS) and Cherchar Abrasion Index values for the fresh and
slightly weathered rock are summarised in Table 3.
Table 3. The main characteristics of the main type of rock encountered during the driving of tunnels A and B
Tunnel A
Tunnel B
Main rock type
Range of rock strength (SI)
71 to 185.9 MPa
75 to 200 MPa
Range of rock strength (TBM)
36.6 to 169.4 MPa
46.5 to 192 MPa
Median rock strength (SI)
151 MPa
140 MPa
Median Rock Strength (TBM)
131.5 MPa
118 MPa
Cherchar Abrasion Index (range)
Average 4.6
3.5 to 4.5
In Table 3, the UCS values for the rock are quoted in relation to the Site Investigation (SI) and as de-
rived from the TBM data (TBM). The quoted values from the site investigation are derived from the
Geotechnical Baseline Report or Geotechnical Engineering Report for the project. The TBM values
were derived from the TBM parameters, using the Colorado School of Mines (CSM) method; the rel-
evant equations are in Rostaimi et al (1996). The values given for the ranges of strength and median
strength are reasonably comparable between those derived from the SI data and from the TBM de-
rived values. Overall, the TBM derived values are slightly lower than the interpreted SI. This is prob-
ably because of a number of factors:
1. The CSM model does not take into account the fracturing of the rock; the effects of fracturing will
be seen as an apparent reduction in strength. This is because it is easier to cut highly fractured
rock than massive rock, for a given UCS value. This will lead to values derived from the CSM
method underestimating the UCS values where the TBM is in highly fractured rock
2. The SI values quoted are for Fresh and Slightly weathered rock. The TBM data probably includes
a significant proportion of moderately weathered rock, which would be weakened by the greater
degree of weathering compared with the Fresh and Slightly weathered rock.
3. There is a bias towards selecting relatively stronger samples rock for testing in the site investiga-
Despite these factors, there is a reasonable degree of comparability in the UCS data between the as-
sessed values from the site investigation and as back-analysed from the TBM data.
3.2 Summary of Tunnel A
The driving of Tunnel A from Ring 150 to Ring 600 is summarized here. Each ring was 1.5m in
length. The face pressure at tunnel axis level was between 3.35 and 4 bar, providing a small overpres-
sure compared with the water pressure where the tunnel was in soil or mixed ground; in rock the face
pressure balanced the water pressure.
The actual ground conditions encountered are summarised in Table 4.
Table 4. Actual ground conditions encountered, Tunnel A
From (Ring No)
To (Ring No)
Ground conditions
Full face of rock grades of granite, except for mixed ground between
R267 and R274, and local dykes of rhyolite and basalt.
Mixed ground (rock and soil grades of weathered granite)
Soil grades of weathered granite
As shown in Table 4, from R322 to R600 the tunnelling was generally in completely decomposed
granite (cdg), a saprolite. However, a number of thin dykes of basalt and rhyolite crossed the tunnel
alignment. The dyke rock within the cdg was generally also completely weathered, although some ev-
idence of basalt rock and corestone boulders was found in samples from the slurry treatment plant. It
was observed at the slurry treatment plant that there were a series of bands of ground that had relative-
ly high clay content within the cdg (clayey cdg). When the TBM was in ‘clayey cdg’, it was found
that there was an increase in the average fines content of the spoil, compared with the remaining cdg.
Evidence of clogging was also noted during interventions, when the face was in the clayey cdg. Many
of the disc cutters damaged in the clayey cdg were flat, particularly the discs mounted in the central
area of the cutterhead.
The calculated values of the Penetration Index for R150 to R600 are plotted in Figure 6; those for
Specific Energy are plotted in Figure 7. The values for the Penetration Index and Specific Energy
were high in the bands of clayey cdg, much higher than in the more typical, granular, cdg. The values
for Specific Energy recorded in the clayey cdg were generally as high, or higher, than recorded when
tunnelling in the granite rock. In contrast, the values recorded when tunnelling in the cdg between the
clayey cdg bands were very much lower. The bands of clayey cdg identified from the Specific Energy
during tunnelling, typically 60m to 120m in thickness, correlated with where increased fines content
was recorded at the slurry treatment plant.
Tunnelling in the mixed ground at the transition from rock to saprolite required a lower Penetration
Index than tunnelling in the rock, but similar or slightly higher values for Specific Energy. The effect
of the various ground conditions on the tunnelling can be assessed in relation to: average instantane-
ous advance rate, the frequency with which discs had to be changed, and the time required for inter-
ventions. These are summarised in Table 5. The values given are averages over all of the rings as-
sessed for a particular ground condition. The times given are average times per metre of tunnel
constructed, and are only for advancing the TBM and interventions. These two activities comprise on-
ly part of the time required for tunnelling, but are the activities that are directly affected by the ground
conditions encountered.
Figure 6. The Penetration Index, R150 to R600 at Tunnel A
Figure 7. Specific Energy, R150 to R600 at Tunnel A
Table 5. Effect on the actual ground conditions encountered on the tunnelling, Tunnel A
Av. Inst. Advance
rate (mm/min)
Discs (m3/disc
Av. Time to
Av. Time, In-
Advance + In-
100% rock
Soil (Clayey
Soil (cdg)
The number of discs changed per cubic metre of ground excavated is presented in Table 5. In order to
arrive at this figure, a simple assumption has been made that the wear of a disc is related to the ground
conditions at the point where the disc is changed. This is imprecise, as the wear on the disc will be the
result of the cumulative wear from the ground conditions since the disc was installed. However, over
long lengths of tunnelling there will be a tendency for more discs to be changed in the more abrasive
ground conditions. Over long lengths of tunnel, the value calculated in this way will be an approxima-
tion of the relative rate of disc wear/damage per metre of tunnel in the different ground conditions.
The time required for interventions was taken from when the TBM stopped advancing, prior to the in-
tervention, to when the TBM resumed advancing. Some time for ring building or other maintenance
may have overlapped with the time for the interventions. The time required was then averaged over
each of the ground conditions identified in Table 5.
There was a reduction in the advance rate in clayey cdg, compared with cdg, of nearly 40%, on aver-
age. On average, there was little difference in the performance of the TBM between rock and mixed
ground. The average advance rate was marginally faster in the mixed ground than the rock, but the av-
erage intervention time was longer in the mixed ground.
3.3 Summary of Tunnel B
The tunnelling for part of the drive, from Ring 1 to Ring 600 is presented, comprising the first 1,080m
of tunnelling. A full face of rock grades of weathered tuff was encountered between R284 and R405,
and locally at R498 and R507. Most of the remaining drive was in saprolite (completely weathered
Tuff), but with local areas of corestone boulders or mixed face conditions, which was defined for this
tunnel as >15% rock in the face.
The actual ground conditions encountered are summarised in Table 6. The assessed ground conditions
were based on all of the information collected before and during tunnelling, and were significantly
different to the conditions anticipated before tunnelling.
Table 6. Actual ground conditions encountered, Tunnel B
From (Ring No)
To (Ring No)
Ground conditions
Completely decomposed Tuff, with some corestone boulders & local ar-
eas of mixed ground
Mixed ground
Rock; locally mixed ground towards the end of the section
Mixed ground
Completely decomposed Tuff, with some corestone boulders
Mixed ground with local area o full face rock at R498 and R507
Completely decomposed Tuff, with some corestone boulders & local ar-
eas of mixed ground
The tunnel was mostly driven in pressurized EPB mode, at up to 3.14 bar face pressure, but open
mode was used between R284 and 405, where the tunnel was generally in a full face of rock. The abil-
ity to maintain EPB pressure in the mixed ground can be contrasted with the problems with maintain-
ing pressure that was reported on the earlier North East Line tunnelling in similar conditions (Shirlaw
2000). This suggests that there were substantial improvements in TBM design and the application of
conditioners, for these ground conditions, since 2000.
Figure 8. Penetration Index at Tunnel B
Figure 9. Specific Energy at Tunnel B
Major areas of mixed ground (MG) and full face rock are indicated in Figures 8 and 9. For clarity,
mixed ground is only shown where 10 rings (18m) of tunnel or more were continuously in mixed
ground. Some locally relatively high values for Penetration Index are evident in Figure 8, such as at
R47 to 52, R164 to R167 and R585 to R591. These represent local areas of mixed face conditions.
Some extreme values of the Penetration Index are shown in Figure 8, of up to 400 MN/m; much high-
er than the typical values of 40 to 60 MN/m recorded in a full face of Tuff bedrock. These values were
also much higher than any value recorded in the slurry drive at Tunnel A. The extreme values for the
Penetration Index generally occurred in mixed ground, close to where the TBM entered or exited the
rock, in EPB mode. These very high values were recorded despite the use of very large quantities of
conditioning agents, typically foam and/or bentonite slurry. In the mixed ground with less than 50%
rock, the consumption of conditioner averaged about 60% of the volume of ground excavated. In
mixed ground with greater than 85% rock, the average consumption of conditioner was 112% of the
volume of the ground excavated.
There were also extreme values of the Specific Energy, up to 400 Mj/m3, just as the TBM approached
or exited full face rock, under EPB pressure. These values were also much greater than the typical
values of 40 to 60 Mj/m3 recorded in a full face of granite rock, in open mode.
The relationship between the ground conditions and the Penetration Index and the Specific Energy
was explored further, by assessing the relationship between the percentage of rock in the face and the
calculated values for that ring of Penetration index and Specific Energy. In the assessment, locations
where there was direct evidence of the actual ground conditions in the face were identified. This evi-
dence came from boreholes located within half a tunnel diameter of the extrados of the tunnel and
from the face maps prepared during interventions. Interventions had been very frequent in the mixed
ground and rock sections of the tunnel, and interventions provided the majority of the information on
the actual ground conditions.
The percentage of rock in the face was plotted against the recorded Penetration Index and Specific
Energy for the relevant ring (Figures 10 and 11). There is considerable scatter in the data. However,
within this scatter there is a clear trend. Up to 50% rock the data for Penetration Index and Specific
Energy increases linearly between the values for 100% soil and 100% rock. The soil and mixed
ground tunnelling with less than 50% rock was all carried out in EPB mode, whereas the rock tunnel-
ling was in open mode. This shows that up to 50% rock, the application of the EPB pressure did not
have a significant effect on tunnelling performance.
Figure 10. Penetration Index plotted against percent-
age of rock in the face, Tunnel B
Figure 11. Specific Energy plotted against percentage
of rock in the face, Tunnel B
When the proportion of rock in the face increased above about 50%, there was a dramatic increase in
both Penetration Index and Specific Energy, well above trend. The peak values occurred just as the
tunnel was transitioning from Mixed Ground conditions to a full face of rock, or vice versa; condi-
tions where a high percentage of rock would occur, but the TBM would still generally be in EPB pres-
sure. Typically, the maximum values in Figures 10 and 11 occur at 85% to 95% rock, with a drop in
value at over 95% rock. Most of the rings with >95% rock were tunnelled in open mode, without EPB
pressure. The extreme values for Penetration Index and Specific Energy recorded in Figures 8 and 9
are not captured in Figures 10 and 11, as there was no intervention or borehole at the relevant rings.
However, it is known that the peak values occurred in EPB mode when the TBM was in a mixed face
comprising mainly rock.
There were extremely frequent interventions where the TBM was in mixed ground conditions with
over 50% rock in the face. High heat was generated when operating in EPB mode with a high percent-
age of rock, with the spoil temperature being measured at over 70OC. In these conditions, the excava-
tion chamber was initially too hot for safe man entry, and the chamber was flushed continuously with
bentonite slurry to cool it down prior to the intervention. Where the machine required cooling, 12
hours to 36 hours of flushing were typically required to achieve safe conditions for man-entry. For
nearly half of the interventions in these conditions, the face was found to be partially caved when the
intervention could safely commence; it is likely that this caving was a result of the long period of
flushing. The partially caved face could generally be maintained in a reasonably stable state, using
compressed air, for short interventions. However, there were some further stability problems during
long interventions. The initial caving of the face probably adversely affected the long term stability of
the face under compressed air.
The additional energy required to advance the TBM under EPB pressure in 50% to 95% rock was
manifesting as heat.
For assessment purposes, the assessed ground conditions along the tunnel were divided up into five
categories, as shown in Table 7.
Table 7. Effect on the tunnelling of the actual ground conditions encountered, Tunnel B
Av. Inst. Advance
rate (mm/min)
(m3/disc re-
Av. Time to
Av. Time, In-
Advance + Interven-
tions, min/m
100% rock
When the rock comprised over 85% of the area of the face, on average interventions were carried out
for just over every 3m of tunnel. Many of these interventions were of very long duration.
It can be seen from the data in Table 7 that the time required to advance the TBM and for the interven-
tions was approximately the same for up to 85% rock, although both values increased significantly
with increasing percentage of rock in the face. However, when the proportion of rock exceeded 85%,
the time required for interventions became extreme, at over 2 days per metre of tunnel driven. In these
conditions, with instantaneous advance rates of typically 2mm to 6mm/minute, and an average of 6
days of intervention time per 3m of tunnel driven, progress was extremely slow. Overall production in
these conditions was about 3m per week, on average. Fortunately, only 1.65% of the total tunnel drive
was in these conditions.
There was one sinkhole recorded over Tunnel B; this occurred after an extremely long intervention in
mixed face conditions with over 85% rock. The time required to recover and restart the tunnelling af-
ter the sinkhole is excluded from the average times given in Table 7.
The rock encountered at Tunnels A and B was approximately comparable, although the granite at
Tunnel A had a slightly higher median strength and was more abrasive than the Tuff at Tunnel B.
Where the two tunnels were in mixed ground, the major difference between Tunnel A, the slurry TBM
drive, and Tunnel B, the EPB drive, was in the performance of the TBMs in conditions comprising
>50% rock. The slurry TBM at Tunnel A passed from rock to soil without significant problems, as
demonstrated by the relatively consistent values for Penetration Index and Specific Energy seen in
Figures 6 and 7. In contrast, once the percentage of rock in the face exceeded 50% the EPB drive at
Tunnel B exhibited rapidly increasing values for both Penetration Index and Specific Energy. These
values increased to extreme values at about 85% rock. The practical effect on EPB tunnelling in these
conditions is demonstrated by the slow average instantaneous advance rate, and number and duration
of interventions recorded. As demonstrated in Table 7, in the most extreme conditions, in a mixed
face with over 85% rock, each metre of tunnelling required on average about 4 hours of TBM advance
and 2 days of intervention time. These values imply an average advance of less than 3m per week, alt-
hough in practice the common pattern was to advance in short bursts of 3m to 6m over a few days, and
then stop for an intervention lasting a week or more. For Tunnel B, it was fortunate that the most on-
erous conditions, 85% to 99% rock, were of limited extent. Other EPB driven tunnels have not been
so fortunate. Shirlaw (2016) provides an example of an EPB driven tunnel which encountered similar
conditions, in weathered granodiorite, over 256m of continuous tunnelling. There were 116 interven-
tions, to replace a total of 513 No 17” discs over that 256m of tunnelling. It took 9.5 months just to
tunnel those 256m, at an average of just over 6m per week; much of the tunnelling was in semi-EPB
mode, as the TBM could generally not be advanced in EPB mode. Another example is the North-
South line of the Bangalore Metro, which also used EPB TBMs in mixed ground of weathered rock
(granitic gneiss which was extremely strong when fresh). One drive of 432m length required 22
months to complete, or an average rate of 4.53m per week; the parallel drive took 12 months, at an av-
erage rate of 8.3m/week (Kenyon 2016).
One of the outward signs of these problems, with tunnelling in mixed face conditions with a high per-
centage of strong rock, is the high temperature of the spoil as it emerges from the screw conveyor.
High spoil temperatures have been recorded on a number of EPB drives other than Tunnel B, includ-
ing a drive in a mixed face of sandstone and soft clay under the Singapore River (Shirlaw et al 2000).
Sometimes, an EPB TBM cannot be advanced in conventional EPB mode in these conditions, as the
cutterhead cannot be turned without breaching the limiting values for the torque. A common expedient,
if this occurs, is to switch to semi-EPB mode (GEO 2014), with the excavation chamber half empty,
and the upper half of the face supported by compressed air. The mode was not used at Tunnel B, but
was used locally for parts of the DTSS tunnelling in weathered granite, and for some tunnelling on the
Circle Line (Venkta & Hoblyn, 2008) in Sandstone of the weathered Jurong Formation.
The problems of high heat and exceptionally high torque during TBM advance appear to be limited to
mixed faces including Strong (or Stronger) and abrasive rocks, when operating in EPB mode, and
where the proportion of rock in the face is >50%. Despite the general trend of using slurry TBMs in
the weathered granite, PTBM tunnelling in the Fort Canning Boulder Bed in Singapore has been en-
tirely with EPB TBMs, on the Circle and Downtown Lines. The Formation comprises about 25% by
volume of Strong or Very Strong and extremely abrasive quartzite boulders in a silty clay matrix
(Shirlaw et al 2003). The continued use of EPB TBMs in these conditions confirms the critical effect
of the percentage of rock in the face.
At Tunnel B, the dramatic increase in the Penetration Index and Specific Energy when the face con-
sisted of more than about 50% rock suggests that the problems are associated with a change from the
spoil being mostly soil, with some chips of rock, to being mostly chips of rock with some soil. The
high heat, high torque and rapid tool wear are suggestive of the spoil clogging in the gap between the
face and the cutterhead (the tool gap). The discs will cut the rock into chips of typically 25mm to
100mm size, based on grading curves from hard rock TBMs (Bruland 1998). Coarse particle clogging
can occur when the maximum size of the chips of rock is more than one third of the size of the tool
gap. The tool gap with a 17” disc is about 150mm, so there is the potential for coarse particle clogging.
As the percentage of rock in the face increases, there will be an increasing proportion of rock chips in
the spoil. This will increase the chance of local clogging. A local area of clogging will allow some of
the contact force applied to the cutterhead to pass through the clogging particles onto the rock in the
face, rather than through the tools. The clogging will also affect the torque, as torque is needed to
break up the clogged particles, which may require the particles to be broken or abraded.
Although this type of problem has been encountered locally in the moderately Strong to Very Strong
Sandstone and meta-Sandstone of the Jurong Formation, it has not been encountered in the weaker
Mudstones and Siltstones of the formation. These weaker rocks are more readily broken or abraded to
break up coarse particles that are clogging.
The support pressure in an EPB TBM operating in full EPB mode is provided through particle to par-
ticle contact within the spoil and onto the face. While under these contact forces, the particles have to
move relative to the cutterhead: from where they are cut to the openings in the cutterhead. Where the
particles are predominantly chips of rock within a narrow slot between the cutterhead and a face of
strong rock, localized clogging can occur, leading to the high values of contact force and torque, but
very slow advance rates, as seen at Tunnel B when the TBM was in a mixed face with a high propor-
tion of rock. In a slurry TBM the face pressure is provided by the slurry, and the cut particles are
transported within the slurry. The slurry TBM at Tunnel A did not experience similar problems in
mixed ground as the EPB at Tunnel B; it is suggested that this was in large measure because of the
different operating principles of slurry and EPB TBMs. The slurry TBM at Tunnel A did experience
some difficulty with clogging in unusually clayey CDG, but this was in relation to fine particle clog-
As discussed above, since 2003 some major project owners in Singapore, including the Land
Transport Authority (for MRT tunnels) and PowerGrid (for cable tunnels) have chosen to specify the
use of slurry TBMs where the tunnel is likely to encounter a significant length of mixed face condi-
tions. The example of Tunnel B, which were constructed after 2010, but not in Singapore, and the
other examples such as the Bangalore N-S metro line, shows that this was a well-founded decision.
The use of semi-EPB mode can often allow an EPB TBM to advance, if it cannot advance in full EPB
mode. In semi-EPB mode only half of the chamber is filled with spoil, reduce the area subject to local
clogging. Also, much of the face pressure is provided by compressed air pressure, rather than by parti-
cle to particle contact, and the spoil is not confined vertically. There are, however, risks in the use of
semi-EPB mode, as the compressed air alone cannot stabilize confined sand layers. This can lead to
loss of ground and large surface settlement; Venkta & Hoblyn (2008) provide a case study on this.
As identified above, there has been a significant improvement since 2000 in EPB tunnelling in mixed
ground conditions. However, further improvements still need to be made to EPB TBMs to improve
their performance in mixed ground conditions comprising soil and strong (or stronger) rock. On the
basis that the problem described above originates in local clogging of coarse particles (chips of rock)
in the tool gap, and manifests as slow advance rates, heat, rapid tool wear and frequent, long interven-
tions, possible improvements could include:
a) Flush with water or (preferably) slurry directly into the tool gap. EPB shields typically include
a number of small nozzles to allow the injection of foam or polymer directly into the tool gap.
Flushing with large volumes of water or bentonite, however, is normally only provided into
the excavation chamber. The direct addition of water or slurry at the tool gap in front of the
cutterhead will provide lubrication to the chips of rock, will assist in cooling, and (for flush-
ing with slurry) will provide clay particles to aid the effectiveness of the foam.
b) Avoid creating unbroken ‘walls’ of cutters which can help to create clogging
c) Minimise the length of travel between the excavated particles and the openings in the cutter-
head, and avoid creating significant areas of the cutterhead with no openings, or openings too
small for the coarser chips of rock to enter
d) Use the largest discs practical: the larger the disc, the more robust it is; also the larger disc
will increase the size of the tool gap, reducing the tendency for coarse particles to clog
e) Plan the compressed air locks and procedures to maximize the effective time that can be
worked during an intervention: Merrie (2009) discusses the benefits of double sets of man-
f) Keep the intervention pressure to the minimum necessary. Shirlaw et al. (2015) discuss the use
of the observational method to allow planned adjustment to the initially estimated compressed
air pressure in Hong Kong
g) Provide a filter cake on the face (GEO 2014) prior to entry, and for long interventions or
where the face is starting to cave or ravel, apply a sprayed membrane (Babendererde & Elsner
However, it is not certain to what degree measures such as those listed above can improve EPB per-
formance in these conditions. For the moment, the choice of a slurry shield where there is significant
tunnelling in mixed ground comprising soil and Strong (or Stronger) rock would appear to be the safer
The use of PTBMs in Singapore has a history stretching back over 30 years. Since 2003 there has
been a general trend towards using slurry shields for tunnelling in the weathered rocks of the Singa-
pore Granite, and, to a lesser extent, the meta-sedimentary rocks of the Jurong Formation. This change
has been a result of project owners specifying the use of slurry TBMs for selected projects, after earli-
er problems with EPB TBMs.
The examples of Tunnels A and B, although not in Singapore, show how an EPB TBM (Tunnel B)
can encounter severe difficulties in mixed ground conditions in the face, comprising greater than 50%
Strong (or stronger) rock. The TBM experienced very slow progress in the worst conditions (>85%
rock, in EPB mode), averaging about 3m per week. Much of the time was spent on interventions, with
very rapid wear and damage to the discs. Similar problems, of very slow progress, heat and frequent
interventions, have been experienced on EPB drives in mixed face conditions on other projects. A
slurry TBM at Tunnel A did not experience the same degree of difficulty in similar mixed face condi-
tions. However, EPB TBMs have been effective in tunnelling in the Fort Canning Boulder Bed, where
the boulders are very strong and extremely abrasive, but the typical proportion of the boulders is about
25% of the ground. This is consistent with the critical effect on the tunnelling of a majority of rock in
the face.
It is postulated that the problems with EPB tunnelling in conditions of greater than 50% rock are re-
lated to local clogging of the chips of rock produced by the TBM cutting the rock, in the gap between
the cutterhead and the face. It should be possible to improve the design and operation EPB TBMs to
improve their performance in these conditions. However, the use of slurry TBMs provides an effective
means of tunnelling in these very difficult conditions.
Babendererde, T. & Elsner, P. (2014) Face support and stabilization in soft ground TBM tunnelling, Under-
ground Singapore 2014, pp 166-176
Balasubramanian, K. (1987) Construction of effluent pipeline using earth pressure balance shield. Proceedings of
5th International Geotechnical Seminar, Case Histories in Soft Clay, NTU, Singapore 1987, pp17-40
Bruland, A. (2000) Hard Rock Tunnel Boring, The Boring Process. Vol 7 of 10, Doctoral Thesis, NTNU
1998 :81
Elias, V.H. and Mizuno, A. (1987) Tunnelling with earth pressure balance shields, Proceedings of 5th Internation-
al Geotechnical Seminar, Case Histories in Soft Clay, NTU, Singapore 1987, pp41-58
Fletcher, C.J.N. (2004) Geology of Site Investigation Boreholes from Hong Kong
GEO (2014). GEO Report No. 298: Ground Control for EPB TBM Tunnelling. Geotechnical Engineering Office,
Hong Kong
Kenyon, P. (2016) Getting Bangalore Metro North-South line back on track. TunnelTalk Jan 2016
Marshall, R.H. & Flanagan, R.F. (2007) Singapore’s Deep Tunnel Sewerage System Experiences and Chal-
lenges. Proc. RETC, Toronto, ed. M. Traylor & J. Townsend publ. SMME, pp1308-1319
Merrie, M. (2009) Trials and Tribulations An Overview of slurry tunnelling in challenging ground conditions.
Underground Singapore 2009, pp12-32
Nakano A., Sahabdeen M. M., Kulaindran A., Seah T. P., (2007) Excavation Management for Slurry TBMs Tun-
nelling under Residential Houses at C853. Underground Singapore 2007 pp38-45
Rostami, J., Ozdemir, L.and Nilsen, B. (1996) Comparison Between CSM and NTH Hard Rock TBM Perfor-
mance Prediction Models, Institute of Shaft Drilling Technology 1996, Las Vegas NV.
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. Invited lecture Proceedings GeoEng2000, Melbourne, Australia, 19-
24 November 2000. Technomic Publishing, PA, USA. Volume 1: 1286-1329
Shirlaw, J.N., Broome, P.B., Changrasegaran, J., Daley, J., Orihara, K., Raju, G.V.R., Tang, S.K., Wong, I.H.,
Wong, K.S., and Kyi Yu. (2003). The Fort Canning Boulder Bed. Underground Singapore 2003. November,
NTU, Singapore, pp388-407
Shirlaw, J.N. & Boone, S. (2005). The risk of very large settlements due to EPB tunnelling. 12th Australian Tun-
nelling Conference, Brisbane, April 2005
Shirlaw, J.N. and Hulme, T.W. (2008) Risk mitigation for slurry TBMs. Tunnels and Tunnelling International,
April 2008, pp26-30
Shirlaw, N. (2015) Pressurised TBM tunnelling in mixed face conditions resulting from tropical weathering of
igneous rock. Int’l Conf. on Tunnel Boring Machines in Difficult Grounds (TBM DiGs), Singapore, 18-20
November 2015
Shirlaw, N., Salisbury, D., Chau, P. and Pang, P.L.R. (2015). The development and application of guidance
documents for ground control for slurry and EPB tunnelling in Hong Kong. Int’l Conf. on Tunnel Boring Ma-
chines in Difficult Grounds (TBM DiGs), Singapore, 18-20 November 2015
Shirlaw, N. (2016). Pressurised TBM tunnelling in mixed face conditions. Hong Kong Tunnelling Society
presentation, January 2016 (available at Researchgate)
Wallis, S. (2004). The good, bad and mixed on the DTSS. Tunnel Talk, April 2004
Wallis, S. (2005). Two extremes in one for Singapore’s DTSS. Tunnel Talk, March 2005
Venkta, R. & Hoblyn, S. (2008). EPB Tunnelling under 2 storey shophouses in mixed face conditions. Inter-
national Conference on Deep Excavations (ICDE) 2008 Singapore
... The use of slurry shields has become common in Singapore and Hong Kong when it is anticipated that weathered granite including a significant proportion of mixed-face tunnelling will be encountered (Shirlaw, 2016b). If EPB TBMs are used, it is typically recommended that an additional active face support system using pressurised slurry is incorporated into the TBM, as developed for the Metro Do Porto tunnelling (Babendererde et al., 2005). ...
Full-text available
Discussion on a case study of EPB tunnelling in weathered granite with boulders
Full-text available
The use of Earth Pressure Balance (EPB) shields is still very modest for tunneling in urban areas in Brazil. This study analyzed and integrated geological data and excavation parameters of a segment of Route 2 (Lot 3) of the Line 5—Lilac (São Paulo Metro), between Campo Belo Station and Jesuíno Maciel Shaft. This segment is 557 m long and represents one of the first experiences on tunneling in transitional mixed face with pressurized shield machines in Brazil. The excavation performed an average advance rate of 7.7 m/day and intercepted different proportions of gneisses, saprolites, and residual soils. The results show how the excavation parameters respond to different geological materials. The sections excavated in mixed transitional ground conditions had several negative effects on the excavation process, similarly to that described in other projects worldwide. Finally, results also suggested the nominal torque was insufficient for excavation of the segment.
Conference Paper
Full-text available
A colluvial deposit, comprising very strong sandstone or quartzite boulders in a hard matrix underlies parts of the central business and commercial areas of Singapore. Foundation, construction, tunnelling and excavations in this deposit pose particular problems; delays and cost overruns during construction are common. Understanding the nature and distribution of the deposit is therefore of considerable value. Although a number of papers have already been published on the nature and extent of the deposit, much previously unpublished information has been obtained in recent years. The information from the previous papers and the new data is synthesised to give an overview of the current understanding of where the deposit has been found, the properties of the boulders and the matrix, and the types of construction work that have been carried out in it.
Conference Paper
Full-text available
Tunnelling using EPB shields generally allows good control of surface settlement. However, there are reports from many projects of occasional, localised, sudden, and very large settlements or sinkholes over EPB driven tunnels. There are a number of readily identifiable risk factors associated with sudden, very large, localised settlements. The relative influence of these risk factors is identified using detailed data from seven major EPB projects with which the authors have been involved. The cases are discussed and illustrated, and high risk areas are identified. Measures to minimise the risk of large settlements are outlined, and issues related to risk assessment are discussed.
Conference Paper
Full-text available
ABSTRACT: Weathering is a process that turns rock into soil. Deep weathering is prevalent in tropical and sub-tropical areas. The resulting sub-surface conditions can be very onerous for tunnelling, with tunnel drives commonly encountering a significant proportion of mixed face conditions, comprising partly rock and partly soil. Problems that have been encountered have included: inability to maintain the face pressure, ground loss, sinkholes, slow rates of tunnelling, rapid tool wear, damage to tools, mixing arms and other parts of the TBM, very frequent and long interventions, clogging and blow-outs. The nature and extent of the problems on any particular tunnel have depended on the type and design of the TBM, the nature of the rock and the proportion of the tunnel in mixed ground. In Singapore this has resulted in a change from mainly EPB to mainly slurry tunnelling in weathered igneous rock; however, predominantly EPB TBMs have been used in weathered sedimentary rock. Information from EPB and slurry TBM drives is used to illustrate the issues involved.
Conference Paper
Full-text available
The Geotechnical Engineering Office (GEO) of the Civil Engineering and Development Department in Hong Kong published GEO Report No. 249, on ground control for slurry TBM tunnelling in 2009. The companion report GEO Report No. 298, for EPB TBM tunnelling, was published in 2014. These reports serve as a basis for the review of geotechnical control submissions made by practitioners on slurry and Earth Pressure Balance (EPB) TBM tunnelling proposals, for consultation with the Hong Kong Special Administrative Region Government (HKSARG). The reports provide a framework for the calculation of the face and tail void grouting pressures required to guard against Ultimate and Serviceability Limit States, under the ground and groundwater conditions commonly encountered in Hong Kong. They also provide guidance for the calculation of compressed air pressures for TBM interventions. The risk of localized settlement (potentially leading ultimately to a sinkhole) and heave (potentially leading to a blow-out) is considered, with guidance on the risk factors involved in such events. The genesis of the reports, and what was included and excluded in terms of content is discussed. The methods outlined in the reports have been applied on a number of completed or ongoing TBM tunnel projects. The methods have generally proven robust, but appear to be slightly conservative, in practice. The experience of working to the reports will be outlined, with comments on issues arising from that experience.
Full-text available
Slurry generated from TBM use in Singapore show that the Code of Practice for tunneling rightly requires risk assessments to be carried out prior to tunneling. There have been a number of incidents of localized loss of ground. There has been the variation in ground conditions and the use of inefficient machines. There is therefore a need to review the geotechnical and topographic data available, recognize its reliability, assess the machinery obtainable and the workforce available, and evaluate the risks that may arise. A comprehensive contingency plan must be in place and practiced so they can be rapidly implemented if the need arises.
At 48km long, the Phase 1 network of the Deep Tunnel Sewerage System (DTSS) in Singapore passes through much of the island's highly variable geology. Conditions vary from soft compressible clays to loose running sands under high ground water pressures. Zones of very hard, highly abrasive, fresh granite to stretches of completely decomposed rock soils, changing in some cases from meter to meter - and with menacing mixed faces in-between - are taxing the skill and abilities of men and machines alike.
Conference Paper
In tropical areas, weathering is common down to several tens of metres, and can occur to depths of over a hundred metres. Tunnels and excavations in such areas are therefore very likely to encounter weathered rock. Weathering involves progressive weakening of the rock, ultimately to a soil-like condition. Despite this soil-like nature, the weathered rock retains many structural features of the original rock. This means that the behaviour may be significantly different from an apparently similar deposited soil. Both the pattern of weathering development and this retained structure are important in understanding the behaviour of weathered rocks. Some of the issues involved in tunnelling and underground excavation in tropically weathered rocks are illustrated with examples of observed performance in four selected rocks found in Hong Kong and Singapore. The four selected rocks consist of two granitic rock formations and two sedimentary rock formations. The development of weathering in these rocks is contrasted, and related to the nature of the fresh rock. The observed behaviour of the weaker and more soil-like weathering grades of the rocks is discussed in terms of the 'Tunnelman's classification'. Particular issues discussed include: mixed face conditions, abrasion, assessment of strength, swelling, relic joints, erosion, hardpans and collapse.