IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

New Austrian Tunnelling Method (NATM) Tunnelling in CentralWan Chai Bypass

Y.C. Lam
Atkins China Ltd., Hong Kong

T. Leung & P. Poon

AECOM Asia Co. Ltd., Hong Kong

L. Ho
Highways Department of HKSAR, Hong Kong

ABSTRACT
The Central-Wan Chai Bypass and Island Eastern Corridor Link (CWB) is a challenging project
as trunk road tunnels with a combined span of 50m mined beneath the existing Cross-Harbour
Tunnel (CHT). An Observational Approach developed upon the concept of New Austrian
Tunnelling Method (NATM) has been adopted in the tunnel design and construction in the
Project.
In this paper, the concept of NATM will be briefly described. It will be followed by a detailed
discussion on how this concept has been applied in the tunnel design and construction with
illustration of numerical analyses by Finite Element Method (FEM). One of the key aspects with
this approach is that the Designer is responsible to closely review the data of construction
monitoring and compare the in-situ monitoring results with the design values. The successful
partnering approach between the Client and the Contractor, which enables the Designer to
optimize the tunnel support system with effective mobilization of inherent ground strength
through deformation, will be presented and recommended for future projects.

1 INTRODUCTION
The works of the entire Central-Wan Chai Bypass and Island Eastern Corridor Link (CWB) project include
the construction of 4.5 km dual three-lane trunk road with 3.7 km section of tunnel starting from Rumsey
Street Flyover at the western side to Island Eastern Corridor at North Point. One of the Contracts, HY/2009/15
entitled Central-Wan Chai Bypass - Tunnel (Causeway Bay Typhoon Shelter Section), was awarded to
China State Construction Engineering Hong Kong Ltd. (CSCE) with a sum of HK$ 5.3 billion. This Contract
includes the construction of 3 contiguous mined tunnels with a total span of 50 m and a length of 160 m at a
depth 26 m below ground. One of the most challenging tasks is to construct this large span tunnel directly
underneath the southern approach ramp of the CHT which was constructed in early 1970s with a series of tiedown anchors. It was assumed that the anchors should be embedded a few metres into bedrock, which is
extremely close to the crown of this new CWB Mined Tunnel. Other sensitive receivers in the vicinity include
the Royal Hong Kong Yacht Club (RHKYC) near the West Portal and the Police Officers Club (POC) near
the East Portal. The layout of the Mined Tunnel is illustrated in Figure 1.

155
1

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Eastbound Tunnel

Westbound Tunnel

Figure 1: General Layout of the Tunnels with Design Support Classes

2 GEOLOGICAL AND HYDROGEOLOGICAL CONDITIONS

2.1 Geology
The tunnel sits on the Hong Kong granite pluton aged Upper Jurassic to Lower Cretaceous, and generally
comprises medium grained and fine grained granite. The boundary between the medium and fine grained
granites is anticipated to be sharp and of sub-vertical contact. Pegmatite pods are common in medium and fine
grained granite boundary. Dykes, as minor intrusions of this area, largely vary in thicknesses. From the GI
data, they mainly consist of feldsparphyric rhyolite and basalt.
The medium grained granite is subject to deep weathering with widely spaced, smooth and rough tectonic
joints as characteristic features. The fine grained granite is porphyritic with phenocrysts of quartz and feldspar,
and the weathering is moderately deep. The jointing is moderately widely spaced with rough sheeting joints.
The superficial geology near the tunnel alignment mostly comprises:
x Saprolite and residual soil derived from in situ weathering of the granitic rocks. Local pockets of residual
soil that is completely decomposed are anticipated;
x Alluvial deposits of the Chek Lap Kok Formation rest on the granitic rocks and are in various states of
weathering, comprising dominantly slightly clayey, silty sand and silty clay;
x Marine deposits with subordinate mud of the Hang Hau Formation on the seabed or underlying the
reclamation adjacent to the mainland, comprising generally sandy, silty clay, and clayey, silty sand, and;
x Reclamation fills of various ages of various materials.
2.1 Geological faults
The NNE trending Wan Chai Gap Fault was expected to intersect the tunnel alignment. It is a strike-slip type
with tens of meters of sinistral displacements. The fault also contains a number of smaller-scale, related faults,
which generally follow the orientation to the main fault zone. Rhyolitic and basalt dykes are common within
the fault zone.

156
2

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

With reference to the geological map, the inferred extent of the rock mass zones considered to have been
disturbed by the fault, are shown on Figure 2. The extent of each is based on an engineering geological
interpretation of the drillhole data, API, published Hong Kong Geology Map and the Quaternary Geology of
Hong Kong and local knowledge of other projects where the same or similar faults have been encountered.
These widths are intended to encompass the rock mass on either side of the faults that may be more
transmissive due to wider discontinuity apertures or significantly increased discontinuity frequency as a result
of the tectonic disturbance. A summary of the fault/fracture zones potentially encountered along the tunnel
alignment and the estimated Q with reference to the available GI are given in Table 2.1 and Table 2.2
respectively.

Figure 2: Geological Layout and Longitudinal Section along Eastbound Tunnel

2.2 Hydrogeology
The tunnel alignment is directly adjacent to the existing seafront. Based on the monitoring records at Quarry
Bay from 1954 to 1999, Table 4 of Part 1 of the Port Works Design Manual suggests that the highest sea level
for a return period of 200 years is +3.60 mPD. However, the design groundwater level (DGWL) was taken as
half a metre above the highest sea level, which was taken as +4.0 mPD, and it was consistent with the
maximum DGWL for tunnel construction in both existing land and permanent reclamation areas under the
Contract requirements.

157
3

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Table 2.1: Summary of geological features along the alignment

No.

Approx. Extent
Along Tunnel
Crown
From
To
(m)
(m)

Characteristics of Geological Features

Part of NE-striking WAN CHAI GAP FAULT ZONE.

F1

4004.0

4017.0

F2 &
F3

4070.9

4083.9

F4 /
Basalt

4147.9

4157.9

18302-III/NP-WCE/T5: Grade III/II fault zone from depth 72 to 76 m and

115.75 to 118.52 m. Closely to medium spaced, rough planar to rough
undulating, chlorite stained and calcite coated joints. RQD = 55 %.
One, NE-trending fault and one NW-trending fault inferred from the DSD HATS
investigation.
LDH14: Grade III/II. Highly fractured zone from depth 9.64 to 14.3 m with
closely and very closely spaced, rough undulating to rough stepped, extremely
narrow, chlorite coated, kaolin (< 2 mm) infilled joints.
NW-strike inferred to be parallel to the Tai Tam Fault Complex.
MDH48: Grade V/IV highly fractured zone from depth 24.2 to 27.45 m. Very
closely to closely spaced, rough undulating to rough planar, limonite and
manganese stained with < 2 mm kaolin infill. RQD < 30 %.
LB11: Grade IV Basalt and Grade IV Granite from depth 27.85 to 31 m and
34.25 to 34.8 m with closely to medium spaced, rough planar and rough stepped,
tight to extremely narrow, iron and manganese stained, chlorite coated joints.
RQD = 50 90 %.

Table 2.2: Estimated Q with reference to pre-construction GI data

Chainage
Predicted Q

Relevant Boreholes

Inferred Faults

4003.96

13.39

MDH39

4003.96

4016.96

1.07

T5, (J24,J25)

F1

4016.96

4042.45

4.32

LDH31-IECL

4042.45

4057.24

7.06

CWBL11

4057.24

4070.87

3.67

LDH14-IECL

4070.87

4083.87

0.31

LDH14, J22a

F2 & F3

4083.87

4088.57

3.67

LDH04-3

4088.57

4100.02

12.28

LB15

4100.02

4111.88

6.46

LB10

4111.88

4125.41

10.19

CWBL14

4125.41

4147.87

13.07

LDH30-IECL

4147.87

4157.87

0.47

LB11, J22

F4/Basalt

From

To

3978.06

158
4

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

3 DESIGN-AS-YOU-GO
3.1 Conventional temporary support design
Structural analyses of the temporary rock support were carried out using two-dimensional, plane-strain finiteelement analyses in computer program Rocscience Phase2. The composite shotcrete with steel ribs/lattice
arch girders was modelled as beam elements surrounded by plane-strain (2D) elements of soil and rock. For
the tunnel sections under general conditions (e.g. other than portals and inferred weakness zones), typical Q
value of 1 was assumed in view of the available GI data. Other parameters were generalized for numerical
assessment, which include:
The shear strength of rock mass was assumed to follow the Hoek-Brown failure criterion. The
corresponding parameters for the Hoek-Brown failure criterion were derived based on the following
assumptions:
For Q = 1.0 or above, UCS of intact rock is 130 MPa;
Q' = 2.5Q (Based on the Qcore assessment, where the value of Q is less than or equal to 1.0, the value of
Jw is either 0.55 or 1.0 and SRF is 2.5. Therefore, multiplying Q with 2.5 results to obtain the value of Q');
Then, the input parameters for derivation of parameters of the Hoek-Brown failure criterions are obtained
using the following relationships:
Geological strength index (GSI)

9 ln Q' + 44

(Hoek et al., 1997)

Rock mass rating (RMR)

9 ln Q + 44

(Bieniawski, 1993)

Youngs modulus, Emass

10 [1.0 (UCS / 100 MPa)]1/3

(Hoek, 2002)

Using an assumed mapped Q-value of 1.0, the following parameters are obtained:
Qdesign

RMR

UCS
(MPa)

GSI

mb

Emass
(GPa)

1.0

2.5

44

130

44

3.925

0.002

0.509

7000

The analyses of rock-structure interaction were carried out separately for the four support classes, SC1,
SC2, SC3A and SC3B with respect to axial, moment and shear capacities of the composite temporary linings.
Each support class was carried out in stages as specified below, which involved temporary load transfer from
innerbound tunnels to the permanent lining of SR8 tunnel.
x
Stage 1: Initial condition prior to excavation
x
Stage 2: Excavation of the top heading of Slip Road 8 (SR8) tunnel outerbound of the top heading of
eastbound (EB) and westbound (WB) tunnels
x
Stage 3: Excavation of the benching of SR8 tunnel and the outerbound of the bottom benching of the
EB and WB tunnels and installation of temporary support
x
Stage 4: Installation of permanent SR8 support
x
Stage 5: Excavation of the innerbound of the top heading of the EB and WB tunnels and installation of
temporary support
x
Stage 6: Excavation of the innerbound of the bottom benching of the EB and WB tunnels and
installation of temporary support
x
Stage 7: Installation of permanent EB and WB support

159
5

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

3.2 Impact assessment on sensitive receivers

In addition to the structural design of the temporary supports, it is important to assess the potential impacts on
nearby sensitive receivers, CHT and its tie-down anchors. To ensure the robustness of the design, no stress
relaxation was conservatively assumed to maximize the potential loads acting to the temporary lining. In
reality, after the tunnel is excavated before installation of support, the ground will be relaxed with stresses redistributed. Hence, an empirical approach proposed by Hoek E. (1999) was adopted to predict the maximum
relaxation (Figure 3).

Figure 3: Predicted relaxation near CHT

Based on the numerical assessment results, it was found that even with 44 % of radial displacement
associated with the ground stress released at the unsupported stage, the estimated settlement at CHT is still
within the acceptable limit. More importantly, the increased axial stress to the tie-down anchors was also
demonstrated to be within the tolerable limit.
3.3 Optimized design under NATM Observational Approach
Instead of adopting the conventional approach with prescribed design support, NATM concepts were
considered in view of the ground conditions and performance of installed temporary supports. In order to
adopt NATM construction approach, close monitoring and timely verification/adjustment of temporary
support design are of paramount importance. Other major design principles have also been followed to
maintain the robustness of temporary support and to ensure construction and public safety.
3.3.1 Design philosophies for NATM
The principles for NATM design can be divided into two functional groups, which are the technical
requirements and the resolutions of external constraints. Regarding the technical requirements, the temporary
support design shall take into account the key aspects, including the geometry, size, excavation sequence,
monitoring systems and all available ground investigation data.
With full compliance to the design standards such as the applicable codes of practice, guidelines, contract
requirements, the NATM design shall appreciate the actual excavation conditions, ground response,
performance of installed supports and last but not the least, the preferred construction sequence.
For the functional group to address other external constraints, which include impact assessment, safety and
environmental issues, contractual and financial constraints, the proper NATM design shall take the right

160
6

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

balance amongst these constraints. For example, contractual requirements with client for optimizing the
design at minimum costs may result in changes to the entire design. In order to successfully apply NATM, it
is important to satisfy the acceptance criteria with comprehensive assessments which can also result with a
cost-effective design.
3.3.2 Key design considerations
By considering the elasto-plastic behavior of the temporary lining, analytical computational models were
developed to estimate earth pressures and displacements of the temporary supports. To take full appreciation
of NATM design, it is important to consider the effects of stress relief ahead of the excavation face and the
construction sequence in stages on the development of temporary load conditions on the temporary lining.
Some key design considerations are summarized below:
(i)
Surrounding rock of the tunnel opening is considered as the main load-carrying component with
full appreciation of homogeneity of the ground, discontinuities, overburden pressures, excavation
patterns and geometry, etc.;
(ii)
Ground response shall be properly assessed in terms of stand-up time and ground-support reaction
curve so that the ground is allowed to deform in a controlled manner, which has been illustrated in
the report published by the Health and Safety Executive (HSE) of UK (Figure 4);
(iii) Water inflow shall be assessed and to provide drain pipes if necessary to allow drainage in
predicted rate with provision of pre-excavation grouting;
(iv) Shotcrete shall be applied, which can be increased in strength if necessary by additional elements
such as thickness, rock dowels, steel mesh, synthetic/steel fibre, so that the time-dependent ground
response and load bearing capacity can be flexibly controlled;
(v)
Excavation shall be carried out in stages to allow controlled deformation and timely support
installation;
(vi) Convergence monitoring of the stresses on the installed temporary support at all stages shall be
recorded and assessed to verify the appropriateness of the NATM design; and
(vii) Due to the importance of stress monitoring, high quality site supervision shall be implemented to
ensure responsive actions appropriately taken on site.

Figure 4: Variation of ground load with deformation during construction (HSE, 1996)

161
7

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

It has been demonstrated that monitoring plays an important role in the NATM construction. The system
should collect qualitative and quantitative data sufficient to meet the needs. As also recommended by HSE, a
flowchart for monitoring and design review during construction has been adopted in construction of Mined
Tunnel (Figure 5). This mechanism was implemented in this tunneling work with comprehensive monitoring
data processing, effective communication link, quality site supervision and responsive design review upon the
ground response and performance of installed tunnel supports.

Figure 5: Monitoring and design review during construction (HSE, 1996)

3.3.3 Adopted risk mitigation measures

Due to the highly sensitive receivers, particularly the CHT as mentioned, it was important to apply mitigation
measures during different stages of the construction. After the Contract awarded in 2010, a series of
additional ground investigation works had been carried out in order to reduce the construction risks by
developing the ground model more precisely, identifying geological and hydrogeological risks more
comprehensively, and providing contingency and precautionary measures more effectively.
Such additional GI works comprised a series of vertical boreholes, inclined boreholes, horizontal
directional coreholes and magnetometry survey for identifying the CHT anchors. The layout of these preconstruction stage GI works is illustrated in Figure 6.

162
8

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Figure 6: Layout of pre-construction ground investigation works

Some key contingency and precautionary measures included i) ground treatment above the portals, ii) preexcavation grouting with pre-determined grout mix, pattern and pressure, iii) barrel and wedge anchors.
With the provision of ground treatment behind the diaphragm walls at the rear of tunnel portals, water
inflow was effectively controlled within the prescribed trigger levels and ensured a safe construction without
causing significant groundwater drawdown or any induced consolidation settlement.
Although an incident of unexpected water inflow due to the installation of inclined spiles was encountered,
an effective pre-excavation grouting work was responsively taken at the site to control the excessive inflow
within a practically manageable time period without causing any noticeable groundwater drawdown in the
proximity or any reportable movement triggering the alert level.
Due to the existence of tie-down anchors underneath the CHT, special precautionary measures were done.
The proactive method by carrying out magnetometry survey was carried out along the horizontal directional
cores. Results indicated that ferromagnetic bodies may be encountered within 1 to 2 m from the horizontal
coreholes. In view of the potential impact induced by the tunnel excavation on these existing anchors, a risk
mitigation plan was proposed with the provision of barrel and wedge anchors. In case of any tie-down anchor
exposed during the excavation, each individual exposed anchor would be properly cut, locked and prestressed with the new barrel and wedge system (Figure 7).

Figure 7: Barrel & wedge anchors

163
9

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Further to those additional GI works, comprehensive monitoring and instrumentation scheme was
proposed and implemented during the construction period (Figure 8). The data were collected periodically,
consolidated systematically, and analyzed numerically for design optimizations with this Design-as-you-go
approach.

Figure 8: Monitoring & instrumentation scheme

3.3.4 Data evaluation & design optimizations

The tunnel temporary support design was initially developed with respect to all the additional GI information.
After the portals were formed, the rockhead level as exposed at each portal is consistent with the design.
However, the granitic rock was found to be more massive and high in strength. After the portal sections were
mined and supported by the prescribed temporary support, more monitoring data were available to optimize
the temporary support designs for Types SC2, SC3A and SC3B mainly in two aspects for enhanced
construction sequence and optimized design support classes subject to better ground reference conditions.
3.3.4.1 Enhancement in construction sequence
As discussed Section 3.1, excavation of the Mined Tunnel involved complicated sequence with respect to
dedicated load transfer in various stages. However, the original schemes were developed upon the limited GI
data without verification of ground behavior and performance of temporary support.
After the completion of the portal excavation, more data including the mapping records, convergence
monitoring, groundwater monitoring and ground surface movement data, had been collected and complied for
design review. One of the major challenges was to adjust the construction sequence to facilitate smooth
construction based on the lesson learnt during the construction of portal sections. A few key amendments
were made with full re-assessment upon the site specific data, which include i) adjustment on the heading
height, ii) lagging distance between different drifts at adjacent excavation faces, iii) pilot tunnel at SR8 tunnel
(Figure 9), and iv) enlarged excavation span of outer section of Westbound tunnel (Figure 10). Each
enhancement enabled a smoother construction and reduced the construction time as a result.

164
10

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Figure 9: Pilot tunnel at Eastern Portal for enabling concurrent excavation of SR8 and Eastbound Tunnels

Figure 10: Enlarged excavation span of outer section of Westbound Tunnel

3.3.4.2 Optimized design support classes
Another proactive approach by optimizing design support classes of SC2, SC3A and SC3B was exercised
which enabled a robust temporary support system to be applied on site safely, promptly and cost effectively
with full assessment upon the actual ground reference conditions and performance of the installed supports at
the previous rounds.
Major optimizations of these support classes include the increase of maximum unsupported distance,
deletion of spile installation, replacement of closely spaced large size steel rib support by fibre reinforced
shotcrete, and supplementary designs with higher Q ranges. The optimized design for different support
classes, namely SC2, SC3A and SC3B are illustrated in Figure 11, Figure 12 and Figure 13 respectively.

165
11

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

Figure 11: Refined design support class SC2

Figure 12: Refined design support class SC3A

Figure 13: Refined design support class SC3B

166
12

IOM3 Hong Kong Branch

Underground Design and Construction Conference 2015

4 CONCLUSIONS
By applying this Design-as-you-go approach, tunnel temporary support was optimized with respect to the
actual ground reference conditions and performance of the installed supports. In general, the actual ground
conditions for Eastbound Tunnel and Westbound Tunnel are favorable with Qmapped averages of 22 and 28
respectively. With the more favorable ground conditions and optimized temporary support system, the average
tunnel excavation rate was significantly improved from less than 0.5 m advancement per day to maximum 2 m
advancement per day and enabled a tunnel breakthrough within 18 months on 15 October 2014 and substantial
completion of Mined Tunnel excavation in April 2015. Indeed, this achievement could only be made with
good trust and effective communication among all key stakeholders including the Employer (Highways
Department), the Engineers Representative (AECOM), the Contractor (China State) and the Contractors
designer (Atkins).
ACKNOWLEDGEMENTS
This paper is published with the kind permission of the Highways Department of HKSAR whom the authors
would like to thank.
REFERENCES
Bieniawski, Z.T. (1993) Classification of rock masses for engineering: the RMR system and future trends, in
Comprehensive Rock Engineering (J.A. Hudson ed.), Pergamon, Vol. 3, pp. 553-73.
Golser J. 1979. Another View of the NATM, Tunnels & Tunnelling, March 1979, Vol. 11, 41.
Grimstad E. and Barton N., 1993. Updating of the Q-system for NMT, International Symposium on Spray
Concrete: Fagernes, Proceedings, 46-66.
Health and Safety Executive (HSE) 1996. Safety of New Austrian Tunnelling Method (NATM) Tunnels, A
Review of Sprayed Concrete Tunnels, HSE Books, 79-137.
Hoek E, Brown ET. Practical estimates of rock mass strength. International Journal of Rock Mechanics and
Mining Sciences, Vol 34, No 8, 1997, pages 1165-1186.
Hoek, E. 1999. Support for very weak rock associated with faults and shear zones Distinguished lecture for
the opening of the International Symposium on Rock Support and Reinforcement Practice in Mining,
Kalgoorlie, Australia, 14-19 March, 1999.
Hoek, E., Carranza-Torres, C. and Corkum, B. (2002) Hoek-Brown failure criterion 2002 edition. Proc.
NARMS-TAC Conference, Toronto, 1: 267-273.
Institution of Civil Engineers (ICE) 1996. Sprayed Concrete Linings (NATM) for Tunnels in Soft Ground,
ICE Design and Practice Guide.
Poon P., Leung T., Chen J. and Ho L., 2015. Construction Challenges of Large Span Highway Tunnel in
Hong Kong Urban Area by Drill-and-Break, Underground Design and Construction Conference
Proceedings 2015, 11 12 September 2015.
Rabcewicz L. 1964. The New Austrian Tunnelling Method, Part One, Water Power, November 1964, 453457; Part Two, Water Power, December 1964, 511-515.
Rabcewicz L. 1965. The New Austrian Tunnelling Method, Part One, Part Three, Water Power, January 1965,
19-24.
Rudolf J., Gall V. and OBrien T., 2010. Past and Present Soft Ground NATM for Tunnel and Shaft
Construction for the Washington, D. C., Metro, 2010 North American Tunneling Conference Proceedings,
2010, 825-835.
Soldata Geophysic, 2013. Geophysical Survey for Anchors Detection using Magnetometry Method, Contract
No. HY/2009/15 Central-Wan Chai Bypass Tunnel (Causeway Bay Typhoon Shelter Section), 1
November 2013.

167
13