Académique Documents
Professionnel Documents
Culture Documents
Wilson W S MOK
Black & Veatch Hong Kong Limited
Maxwell K W MAK
Drainage Services Department
Felix H T POON
Black & Veatch Hong Kong Limited
Introduction
Pipejacking techniques have been commonly adopted in the urban areas
of Hong Kong for the installation of pipelines, such as sewers, drains
and watermains, in recent years. Problems such as excessive ground
settlements, out-of tolerance in line and level, frequent change of disc
cutters, and malfunction of the TBM are often encountered during the
course of works, affecting the adjacent utilities, services, structures, roads
and footways, and more importantly, the daily production.
In the past, many observations on the pipejacking techniques were
based on overseas projects, with little discussions on the local practices.
It is hoped that, through the successful completion of DSD Contract No
DC/2000/11 Wan Chai East and North Point Sewerage Trunk Sewers,
practitioners can gain better understanding of such techniques on the
ground conditions in Hong Kong.
This paper discusses the performance of the pipejacking works observed
for the 4 km long trunk sewers, completed in 21 shallow and deep
drives, under different depths and ground conditions, by the 4 nos slurry
operated TBMs, in two of the most busiest urban areas in the Hong
Kong Island. Analyses of the measured data and comparisons with the
predicted values, with regard to ground behaviours and responses, are
given. Improvements proposed, and different options of TBM driving
to suit different site constraints, and the need in programming, are
highlighted. Recommendations, based on the lessons learned and
experiences gained, are also made.
and the stalled TBM due to obstruction in a drive which resulted in the
requirement of a rescue shaft for changing the cutting wheel before the
resumption of the works.
TBMs also performed signicantly different in excavation time even in
similar ground conditions and curvature. The drives in soft ground usually
had better excavation time than those in mixed or hard ground, with
the highest 77% and the lowest 21%. The presence of a high content
of hard materials sometimes caused damage to the TBM. Frequent
repair and maintenance was often required, signicantly affecting the
TBM performance. However, there was no indication that the straight
alignment had better excavation time and less downtime than the
curved alignment.
For classication purposes, the soft ground is dened as a ground made
of clayey, silty and/or sandy materials such as marine deposits and
alluvium; the mixed ground generally consists of soft materials with rock
in part of the excavation face; the hard ground is a ground consisting
of rock in the full excavation face. Such classications were based on
visual inspection of the materials collected in the muck tank after the
completion of jacking each pipe.
Jacking Force, Rate of Penetration and Friction Force
The average jacking force required to tunnel through lled ground
for shallow drives ranged from 70 to 220 T, with an average rate of
penetration of 110 to 175 mm/min. This rate was reduced to 70 to
90 mm/min when hard materials were encountered.
For deep drives in mixed to hard ground, the average jacking force
was generally in the range of 290 to 480 T, giving an average rate
of penetration of 30 to 90 mm/min. In soft ground such as marine
clay or alluvial deposits, the rate of penetration could be as high as
350 mm/min under a jacking force of 300 T. A tunnel through a full
face of moderately to slightly decomposed granitic (M/SDG) bedrock had
recorded a penetration of 18 mm/min under a consistent jacking force
of 300 to 330 T. The average rate of penetration versus the average
jacking force for different drives is shown in Fig 3.
There was no indication that a higher average jacking force had been
required for curved drives in soft ground, although a jacking force of more
than 750 T had been used at certain locations of the 404 m S-curve
alignment and some of the curved alignments. A higher jacking force
was usually required to push the pipeline after the TBM had stopped in
ground for some time for repair or changing of disc cutters.
The average friction resistance derived from the jacking force on the
drives ranged from 1.2 to 4.8 kN/m2 for clayey materials such as marine
deposits, from 1.9 to 2.8 kN/m2 for sandy materials such as alluvium and
31
TRANSACTIONS PAPER
Drive
Radius (m)
Length (m)
Ground
Conditions
Excavation
Pipe
Positioning
TBM
Downtime
Survey
Check
Slurry
Mixing
Change of
Cutters
Others
245
145
404
370
220
170
58
Soft
Soft
Soft, mixed & hard
Mixed to hard
Soft, mixed & hard
Hard
Soft
Average
62
61
21
45
60
44
39
42
14
24
7
9
9
8
33
11
12
0
52
31
5
28
22
30
7
9
5
4
18
5
6
6
5
6
4
0
0
0
0
2
0
0
11
11
8
15
0
9
0
0
0
0
0
0
0
0
160
40
165
185
150
60
37
17
30
72
21
32
39
20
14
20
9
17
21
17
29
60
14
0
58
47
27
6
9
31
10
4
0
12
8
0
5
9
0
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Soft
77
14
Soft
Soft, mixed and hard
Soft, mixed and hard
Soft with little mixed
Soft, mixed and hard
Soft
Soft, mixed and hard
Average
14
27
11
26
19
28
35
18
36
19
5
41
9
45
12
12
39
4
4
13
10
7
3
7
6
3
1
18
7
18
3
4
5
13
6
2
2
2
9
6
0
34
8
0
53
0
38
20
0
0
65
0
0
0
0
33
107
148
268
307
168
225
210
33
Drive
Diameter
(mm)
Curvature (m)
Total
Length
(m)
5.9 - 5.4
365
150
All soft
Depth to
Invert (m)
% Out of Tolerance
(Contract)
Line
(50 mm)
Level
(35 mm)
1200
NP2 - NP3
1200
5.9 - 6.6
Straight
60
24
WC5 - WC4
1200
5.6 - 6.6
750
160
21
14
WC5 - WC5A
1200
5.6 - 4.7
Straight
40
30 m soft, 10 m mixed
11
22
WC5 - WC7
1200
5.6 - 5.0
340
165
All soft
WC8 - WC7
1200
4.8 - 5.0
405
185
All soft
4
Average Value:
6.5
0
6.3
1800
15.0 - 15.8
Straight
58
All soft
NP5 - NP3
1800
14.2 - 13.7
404
35
19
NP5 - NP7
1800
14.2 - 15.1
Straight-500-Straight
370
17
37
NP8 - NP7
1800
15.4 - 15.1
Straight
220
47
26
NP8 - NP9
1800
15.4 - 15.6
Straight
170
All hard
27
WC2 - WC1
1800
14.0 - 16.5
1500
245
All soft
20
14
WC2 - WC4
1800
7.6 - 6.6
145
All soft
Average Value:
42
21
26.7
24.0
5.0
0.0
600
4.8 - 3.9
Straight
53
All soft
Average Value:
1800
15.8 - 16.2
Straight
307
NP10 - NP9
1800
15.8 - 15.6
800
168
22
41
NP12 - NP11
1800
16.5 - 16.2
1750
268
21
43
NP12 - NP13
1800
16.5 - 16.4
Straight
148
NP12 - PS1
1800
16.5 - 17.0
Straight
107
All soft
57
43
NP14 - NP13
1800
16.0 - 16.4
1185
210
73
29
NP14 - NP15
1800
16.0 - 16.8
1965
225
All soft
Average Value:
Average Value of All Drives:
13
23.5
21.3
21.2
19.2
The average daily production rate for each pipejacking drive versus the
drive length, the maximum daily production rate versus the drive length
and the average contract production rate/the average production rate of
each drive versus the drive length are shown in Figs 5 to 7. The overall
production for each pipejacking drive, completed by the four TBMs, and
for each TBM, is shown in Figs 8 to 10 and 11 respectively.
Replacement of Disc Cutters
A total of 254 nos steel alloy disc cutters were used for the 21 drives,
averaging about 0.5 set of cutters for each shallow drive and 1.5 sets for
each deep drive. A disc cutter had to be replaced when it was damaged
or had a wear of about 12 - 15 mm. Based on the records, the damage
of some of the disc cutters was caused by the dynamic impact of the
TBM, as a result of the sudden change in ground conditions from soft
to hard. The damage also occurred in hard driving under high jacking
forces. It was observed that the damaged disc cutters generally appeared
in the form of breakage of part of the disc, fractured rings and distortion
in the bearing system.
No discs had to be replaced for the Herrenknecht AVN 1200TC TBM
during the course of each drive, but a number of drives required the
replacement of some disc cutters before commencement.
For the Lovat mts 2000 TBM, no replacement of disc cutters was required
for the drives in soft ground. However, the average replacement frequency
for each disc cutter, in the range of 0.4 - 3.3 times, was recorded in
35
Figure 10 Overall Production of Each Drive Using Lavat mts 2000 TBM
highly variable ground made up of soft, mixed and hard materials, for
each drive.
The Herrenknecht AVN 1800T TBM also did not require any replacement
of disc cutters for the drives in soft ground. However, it showed an
average frequency of 1 - 3 times for each disc cutter in highly variable
ground, and 1.9 times in hard ground.
Gauge disc cutters usually required more frequent replacement than centre
disc cutters. This is because, during each rotation of the cutting wheel,
the gauge disc cutters have to travel a longer path for excavation, thus
causing higher consumption. For the Herrenknecht AVN 1800T TBM, the
cutters were replaced in a ratio of 1 (for centre discs) to 2.5 (for gauge
discs), while a 1 to 4 ratio was recorded for the Lovat mts 2000 TBM.
Depending on the ground conditions encountered, particularly the location,
thickness and hardness of materials, some of the disc cutters had to be
replaced again after driving for a short length. This situation occurred in
some long curved drives with highly variable ground conditions and the
replacement of disc cutters in a 4m drive length had been experienced
in the worst case. The replacement of more than 30 nos disc cutters
during the course of works in three drives had been recorded.
Figure 9 Overall Production of Each Drive Using Herrenknecht AVN
1800T TBM
36
For effective monitoring, a numbering system was adopted for the disc
cutters in the cutting wheel, with a number assigned to each disc cutter
Average
Service
Distance
Hard
Mixed
Soft
Herrenknecht
AVN 600 TBM
53 m
19%
38%
43%
Herrenknecht
AVN 1200TC TBM
27
254 m
6%
11%
83%
Herrenknecht
AVN 1800T TBM
153
115 m
30%
19%
51%
69
238 m
15%
10%
75%
TBM
Ground Composition
The transverse settlement trough width for each drive was back calculated
from the measured data. It was found that the trough widths ranged
from 18 to 30 m for the shallow drives, with an average of 19 m, and
from 30 to 60 m for the deep drives, with an average of 35 m. Except
for one drive, these troughs were all atter than those predicted, with
an average of 8 m and 24 m for shallow and deep drives respectively.
This was because although the settlement above the centre line of the
pipeline was less than the predicted value, the settlements at the left
and right hand sides were greater than the predicted value. The average
volume loss for the deep drives, 3.97%, was less than the predicted
4.46%, while for the shallow drives, the average volume loss, 4.86%,
was greater than the predicted 3.15%. However, it is interesting to note
that although the average volume loss was greater than that predicted
for shallow drives, the measured settlements above the centre line of the
pipeline were generally less than the predicted values as the widths of
settlement troughs were underestimated. Besides the overcut of the TBM
assumed in the calculation, other ground losses such as those discussed
in the following paragraphs may also have certain contributions to such
phenomenon.
A number of drives for which the measured ground settlement exceeded
the predicted value can be attributed to no slurry having been injected
to balance the water pressure in the excavated face as a result of
sudden change in ground conditions not immediately noticeable to the
TBM operator. In some cases, the presence of hard materials in part of
the excavation face caused face loss due to the adjacent soft materials
moving faster into the excavation chamber of the TBM, resulting in
more settlement.
It was observed that large settlements usually occurred during hard and
prolonged driving as a result of obstruction, the presence of existing
underground voids, and the migration of the surrounding soil into the
receiving shaft due to the failure of the treated soil block outside the
receiving eye, when the TBM was approaching. Large settlement was
also found at the location beyond the treated soil block outside the
jacking shaft when the complete TBM had not been pushed into the
ground and started to inject slurry to balance the total horizontal soil
and water pressure in the excavation face.
There were some occasions that voids were found at locations of signicant
settlement detected by the monitoring markers. A review on the records
however did not indicate any abnormality in the pipejacking works, such
as over-excavation or excessive loss of slurry. It was suspected that such
voids existed in the ground for some time and the pipejacking works
just triggered the incidents. The voids were then backlled with sandy
materials or grout before the resumption of works.
The results of ground settlement monitoring for a pipejacking drive, that
illustrated the presence of voids below the concrete carriageway, are
shown in Fig 13, and the condition of a void is shown in Fig 14.
For almost all the drives, ground settlement was negligible at a distance
beyond half of the depth of the jacked pipeline ahead of the TBM. The
37
Drive
Diameter
(mm)
Depth to
Invert (m)
Total
Length
(m)
Average
Overburden
SPTN Value
at Shaft
Location
53
11
All soft
Measured Surface
Settlement (mm)
Min
Max
Average
Predicted
Surface
Settlement
(mm)
12
11.0
600
4.8 - 3.9
5
Herrenknecht AVN 1200TC TBM
NP2 - NP1
1200
5.9 - 5.4
150
16
All soft
61
21
14.7
NP2 - NP3
1200
5.9 - 6.6
60
30 m soft, 10 m mixed
and 20 m hard
26
13.7
WC5 - WC4
1200
5.6 - 6.6
160
12
13.2
WC5 - WC7
1200
5.6 - 5.0
165
11
All soft
35
15
16.2
WC5 - WC5A
1200
5.6 - 4.7
40
17
16.0
WC8 - WC7
1200
4.8 - 5.0
185
10
All soft
24
13
16.2
12.2
Herrenknecht AVN 1800T TBM
WC2 - WC1
1800
14.0 - 16.5
245
All soft
21
18.6
WC2 - WC4
1800
7.6 - 6.6
145
All soft
31
11
40.0
NP5 - NP3
1800
14.2 - 13.7
404
32
42
18.1
NP5 - NP7
1800
14.2 - 15.1
370
26
10
18.1
NP8 - NP7
1800
15.4 - 15.1
220
39
17.0
NP8 - NP9
1800
15.4 - 15.6
170
39
All hard
26
16.9
NP15 - NP16
1800
15.0 - 15.8
58
18
All soft
10
71
33
16.7
6.2
Lovat mts 2000 TBM
NP10 - NP9
1800
15.8 - 15.6
168
11
63
25
16.9
NP10 - NP11
1800
15.8 - 16.2
307
10
28
13
16.5
NP12 - NP11
1800
16.5 - 16.2
268
12
54
23
16.0
NP12 - PS1
1800
16.5 - 17.0
107
10
All soft
52
26
15.6
NP12 - NP13
1800
16.5 - 16.4
148
12
24
13
15.9
NP14 - NP13
1800
16.0 - 16.4
210
11
38 m soft, 53 m mixed
and 119 m hard
16.2
NP14 - NP15
1800
16.0 - 15.8
225
14
All soft
18
11
16.5
15.5
11.0
settlement, however, developed rapidly after the TBM had passed through.
An average of 74% and 77% of the settlement occurred during and
immediately after the completion of pipejacking works for the shallow
and deep drives respectively. However, due to the time required for
dismantling the xtures inside the completed pipeline, such as slurry
charge and discharge pipes, electricity cables, water hose, ventilation
hose, and the automatic lubrication system, further settlement developed
as a result of the tendency of the ground to close in the annulus of
the pipeline before the grouting material could be applied. It is noted
that the settlement for shallow drives (which were mainly carried out in
sandy ground) usually stopped within 3 - 4 weeks after the completion of
pipejacking works. For deep drives in highly variable ground, particularly
with clayey materials, the settlement would continue slowly over a much
38
39
This is the traditional method, in that the pipeline will only be jacked
after the completion of the two temporary shafts at each end. This
would ensure the watertightness between the permanent shaft and the
completed pipeline.
With Completion of Intermediate Temporary Shaft in Advance (With
3 or More Shafts in a Line) (Fig 20)
This would require the construction of the receiving and launching eyes
at the piles of the temporary shaft before the jacked pipeline passes
through. The condition of the TBM can be inspected in the shaft and the
damaged disc cutters changed before the pipeline is pushed forward.
The shaft is then ooded with water to balance the groundwater pressure
inside and outside, and a double layer rubber seal is installed at both
eyes to minimise the loss of lubricant and migration of the nes in the
surrounding soil when the jacking continues. A wooden planking system
is provided on top of the pipeline to prevent oatation. It is, however,
noted that in so doing, the friction resistance on the pipes would be
increased. Alternatively, the temporary shaft could be temporarily
backlled for such purposes.
The benet of using this option is the exibility of programming which
could make the re-sequencing of works easier should any site problems
40
existing utilities/services to give room. This would also give the option to
move the shaft location along the completed pipeline if any unexpected
situations arise; however, there may be difculty to effectively treat the
ground at the window locations.
With Completion of Permanent Shaft(s) in Intermediate Temporary
Shafts and Receiving Shaft in Advance (With 3 or More Shafts in a
Line) (Fig 22)
The reinforced concrete structure for the permanent shaft is completed
rst, with the receiving and launching openings prexed and formed.
A mass concrete lling is then provided across the shaft to conne the
alignment of the jacked pipeline and prevent the escape of lubricant when
the pipeline is passing through. This could also enable the inspection of
the condition of the disc cutters and the necessary replacement.
The gap between the temporary and permanent shafts at the opening
locations also needs to be sealed up by concrete to prevent ingress of
water causing drawdown outside the temporary shaft and then attracting
unnecessary ground settlement.
The speed of the TBM, when close to the temporary shaft, is to be
reduced and the position of the TBM is double checked and adjusted,
as necessary, before the TBM is pushed in. The jacking force and the
speed of driving is controlled in such a manner that the vibration induced
(generally less than 25 mm/sec based on the measured data on three
drives) will not damage the permanent shaft.
Upon the completion of the pipejacking works, the gap between the
permanent shaft and the pipeline (with Hi-rib, which is a special type
of construction joint material, provided to increase bonding) is properly
sealed up by non-shrinkage epoxy grout.
For the case in the receiving shaft, the internal size of the permanent
shaft must be large enough to allow the removal of the longest section
of the TBM, and the internal xtures and top slab of the shaft are to be
completed later on.
This option can be adopted effectively at shaft locations where public
complaints in respect of lengthy occupation of works area causing loss
of business and affecting environments are a major problem. Such would
allow the construction of the permanent shaft immediately after the
completion of the intermediate shaft, the reinstatement of the occupied
area (except for the receiving shaft location where the top slab of the
permanent shaft is to be placed after the retrieval of the TBM) and the
release of the site to the public, at a much earlier stage. There is no
restriction on the number of intermediate shafts that the TBM and pipeline
can pass through. However, typical TBMs could generally operate 4 nos
intermediate jacking stations. Therefore, the TBM should be specially
designed for long drives such that sufcient nos of intermediate jacking
stations can be installed and operated, to account for the anticipated
ground conditions and the length of driving.
Conclusions
Certain observations and conclusions may be drawn from the experiences
gained on this contract.
1)
Figure 21 Pipejacking through an Intermediate Shaft without
Completion of Temporary Works in Advance
The average production rate for the drives completed by the 4 nos
TBMs ranged from 2.5 to 14.7 m/day. For planning purposes, a
production rate of 4 m/day is considered to be a reasonable lower
41
More than 20% of the deep sewers exceeded the contract tolerance
for line and level, despite the fact that the excavation face had been
conditioned by slurry and sometimes by the addition of polymer,
as necessary. In consideration of the steering capacity of the TBMs
available in the market, the allowable angular deection of the
jacking pipes, and the procedure of correcting the out-of-tolerance,
a tolerance of 50 mm for line and 35 mm for level may not be
applicable to all Hong Kong soils for pipejacking works. Therefore,
the respective tolerance should only be used with extreme care,
with adjustment as necessary.
75 mm for line and 50 mm for level could be a more realistic
tolerance that the TBM deployed can practically control the alignment
of pipeline in the ground likely to be encountered.
3)
4)
5)
6)
7)
8)
9)
42
10) Four options of TBM driving had been successfully adopted for
different pipejacking drives (with both straight and curved sections),
under different ground conditions. This gives clients and contractors
exibility in designing and programming the works, particularly
when site constraints are encountered.
11) Based on the problems encountered and the lessons learned, certain
areas for improvements are suggested. It is believed that, through
the proper implementation of such improvements, the performance
of pipejacking works could be greatly enhanced.
Acknowledgements
The authors wish to express their gratitude to the Drainage Services
Department of the Government of Hong Kong SAR, for permission to
extract the materials from the respective Project, to publish this paper.
The assistance provided by Mr Derek Arnold, Director of Black & Veatch
Hong Kong Limited, and Mr K C Leung and Mr Ivan S W Wong of the
Resident Site Staff in analysing the eld data, are also appreciated. Special
acknowledgement is given to Herrenknecht AG for his permission to
extract the photographs from technical brochures in making Fig 15.
References
1.
2.
3.
4.
5.
6.
7.
Thomson, J., Pipejacking and Microtunnelling, pp. 182, Black A&P., U.K.,
(1993).
The British Tunnelling Society and The Institution of Civil Engineers, Specication
for Tunnelling, pp. 36, 86 - 88 & 106, Thomas Telford, U.K., (2000).
General Specication for Civil Engineering Works, Volume 1, pp. 139 -140,
Civil Engineering Department, Government of the Hong Kong SAR, Hong Kong
(1992).
Field data of DSD Contract No. DC/95/05 Central, Western and Wan Chai
West Sewerage Trunk Sewers.
Field data of DSD Contract No. DC/98/06 Aberdeen and Ap Lei Chau Trunk
Sewers.
Mok, W. W.S., Performance of Trenchless Techniques for Sewer Construction
in Hong Kong, HKIE Transactions, Volume 9, No. 1, pp. 51 - 56, The Hong
Kong Institution of Engineers, Hong Kong (2002).
OReilly, M.P. and New, B.M., Settlement above Tunnels in the United Kingdom,
Their Magnitude and Prediction, pp.173-181, Proceedings Tunnelling 82
Symposium, Institute of Mining and Metallurgy, U.K. (1982).
43