You are on page 1of 9

Tunnels and Underground Cities: Engineering and Innovation meet Archaeology,

Architecture and Art – Peila, Viggiani & Celestino (Eds)


© 2019 Taylor & Francis Group, London, ISBN 978-1-138-38865-9

Design and construction of large span tunnels and caverns in


Sydney, Australia

D.A.F. Oliveira
Jacobs Engineering Group and University of Wollongong, Australia

ABSTRACT: Increased demand to future-proof tunnel projects with respect to traffic has led
to the proposal of some very large tunnel spans in recent road tunnel projects in Sydney. For
example, four lane tunnels are currently under construction in Sydney with mined spans of
approximately 20 m and Y-junction caverns exceeding 30 m spans, all with a requirement for
100-year design life. As these spans are unprecedented in Australian civil tunnels, a direct com-
parison with local experience is not possible and simple extrapolation of precedent designs, not
necessarily adequate. This paper intends to present and discuss how recent designs that focus on
first principles and the basic objectives of rock reinforcement have overcome the challenges of
these designs, satisfied codes and standards requirements but at the same time provided savings
with respect to ground support. The key to the design involved understanding the key failure
mechanisms that needs to be addressed, its relationship with the different actions of rock bolt-
ing, i.e. suspension/anchorage and/or rock reinforcement and what could be considered accept-
able for design.

1 INTRODUCTION

With rapid development of cities, it is crucial that the use of the underground space is made
efficiently with projects that can cater for the needs of the population for several decades. This
has led to an increased demand to future-proof tunnel projects with respect to traffic which
resulted in the proposal of some very large spans in recent road tunnel projects currently in
construction in Australia all with the requirement for 100-year design life.
For example, several kilometers of four lane tunnels are currently under construction in
Sydney with mined spans of approximately 20 m. Such spans had only been experienced in
localized excavations in widened sections such as breakdown bays and Y-junction caverns but
not for long lengths of tunneling. In addition, the Y-junction caverns now required for these
tunnels are also unprecedented for road tunnels in Australia with spans reaching 31 m and
exceeding experience in Australia which include the Eastern Distributor in Sydney (24 m) and
both Kedron (26 m) and Lutwyche (27 m) caverns of the Airport Link tunnels in Brisbane.
These large span road tunnels are currently in construction for the New M5 and M4-M5
Link tunnels part of the infrastructure project known as WestConnex. The WestConnex project
is a 33-kilometre underground motorway currently being constructed in Sydney’s Inner West
(Figure 1).
As these excavation spans are unprecedented in Australian civil tunneling, a direct compari-
son with local experience is not possible particularly considering the semi-flat roof tunnels typic-
ally excavated in Sydney. Although simple extrapolation of precedent designs could potentially
provide a solution, two risks arise: (1) the extrapolation based on different excavation shapes
that do not necessarily target the actual failure mechanisms involved in the excavation of such
larger spans; and (2) provide uneconomical solutions. International experience could certainly
be used but adequate design justifications and analysis would still have to be provided to verify
its application locally.

5981
Figure 1. WestConnex Motorway.

The search for solutions to new problems often target innovation. However, considering
that the new challenge described above in fact involves an “old” problem but at a larger scale,
it is considered appropriate to review the basic design assumptions to find robust solutions in
more fundamental design principles. As a result, this paper intends to present and discuss how
a design that focus on the basic objectives of rock reinforcement may allow for a better under-
standing of the design requirements and still provide savings with respect to ground support.
The key to the design involves understanding the failure mechanism that needs to be
addressed, its relationship with the different actions of rock bolting, i.e. suspension/anchorage
and/or rock reinforcement and what could be acceptable.

2 PRECEDENT DESIGN

Figure 2 presents a comparison between span and bolt length for several projects in Australia.
Significantly experience is observed for tunnels under 20 m span with most of the projects
in Sydney, i.e. within similar geology (Hakwesbury Sandstone). In addition, a reasonable
number are road tunnels with similar construction methodology, i.e. roadheaders, and excava-
tion profile. For these smaller span tunnels, the bolt length typically follows a power curve
with respect to the span that if assumed for tunnels of approximately 20 m span, a bolts
length of approximately 6 m would correspond to the precedent design.

Figure 2. Span versus bolt length for several projects in Australia.

5982
Figure 3. Typical cavern geometry proposed in Westconnex.

For caverns approaching 30 m span, the precedent bolt length would seem to be approxi-
mately 8 m long. However, it is important to note that very few data correspond to tunnels in
Sydney beyond approximately 18 m span, i.e. only the Eastern Distributor cavern with 9 m
long bolts up to a span of 24 m and Lane Cove Tunnel with 7 m bolts for a 22 m span excava-
tion. The other projects are essentially outside Sydney, namely the Poatina Power Station
cavern in Tasmania with 7 m bolts for an approximately 29m span excavation, the Kedron
cavern in Brisbane with 7 m bolts for a 26 m span and the Lutwyche cavern with 6 m bolts for
an approximately 26 m excavation.
The main difference between these larger excavations outside Sydney and the large span
tunnels excavated in Sydney is that they are all fully arched structures where the tunnels in
Sydney are semi-flat roof tunnels, i.e. very low arch tunnels (Figure 3). Stress arching and
combined use of passive shotcrete support in arched excavations may allow for shorter bolts
than would have been used in semi-flat-roofed tunnels in Sydney. One of the reasons is also
associated with the precedent design and underlying philosophy for flat-roofed tunnels in
Sydney which would generally involve selecting a roof beam thickness based on a level of
deflection deemed acceptable, typically between 15 mm and 20 mm, which would in turn dic-
tate the rock bolt length (Bertuzzi and Pells 2002). As a result, this design approach typically
results in longer rock bolts due to the larger deflections expected in flat-roofed tunnels than in
arched excavations.
As a result, the use of the above precedent design in Australia for the design of the current
large tunnels needs to be assessed with care.

3 FAILURE MECHANISM

A fundamental step in the design of any ground support is the understanding of the geome-
chanical behavior of the structure and how it would ultimately collapse. In other words, the
failure mechanism ultimately dictates the required ground support.
An analogy can be made, for instance, with concrete columns that are subjected to both
axial and flexural loads. An increase in the amount steel reinforcement would provide small
benefit if the primary mode of failure is compression due to high axial loads whereas an
increase in concrete strength or cross-sectional area would likely be required. On the other
hand, an increase in steel reinforcement would likely be the required solution if the main fail-
ure mechanism is flexural.
To confirm the likely failure mechanism associated with such large span excavations, the
construction of the proposed large tunnels in Sydney was simulated using a Distinct Element
Method code, UDEC (ITASCA), and assuming an unsupported full-face excavation. A typ-
ical ground condition observed in Sydney Hawkesbury Sandstone is assumed with Sandstone

5983
Class I/II (SS-I/II) but with relatively adverse thin laminations (SS-II/III) present in the imme-
diate tunnel roof as shown in Figure 4. For the Sydney Rock Classification, the reader is
referred to Pells et al (1998). However, as an immediate simple comparison, the condition
shown in Figure 4 would have a Q-value greater than 4.
The ground behavior was assessed by developing a Ground Reaction curve similarly as in a
Convergence-Confinement approach where the excavation perimeter tractions are reduced
from their initial condition (no excavation or support stress equivalent to the in-situ condi-
tion) to a full excavation (0% support). The failure mechanism and ground reaction curves are
presented in Figure 5.
It can be observed that the failure mechanism is primarily associated with buckling of the
rock beds similar to a “voussoir rock beam” mechanism as often considered in the design of
tunnel support in Sydney (Bertuzzi and Pells 2002). The ground reaction curve indicates that
at a stress level equivalent to approximately 12–14% of the original in-situ stress the rock mass
behavior is within its non-linear range primarily due to yielding along the bedding partings.
At a stress equivalent to approximately 9–10%, the main collapse mechanism is initiated, and
crown sagging becomes more pronounced and asymptotic. The collapse mechanism develops
approximately the following process:
• Partial loss of abutment support due to corner wedges formed by arched roof
• gradual development of bedding plane shearing and slip with inward movements from
abutments
• bedding separation in the mid-span
• coalescence of bedding slip from abutment towards central zone above crown
• buckling of rock beds with propagation towards stiffer and more competent beds (if avail-
able) potentially leading to formation of a chimney type collapse.

Figure 4. Ground conditions assumed.

Figure 5. Collapse of unsupported cavern and associated Ground Reaction Curve.

5984
4 DESIGN PHYLOSOPHY AND SUPPORT STRATEGY

Once the main failure mechanism has been confirmed, the appropriate support interaction
needs to be selected. In general, ground support in underground excavation can be divided
into three main primary functions as illustrated in Figure 6: (a) rock reinforcement; (b) rock
suspension or hold action and (c) surface retention/support.
Rock reinforcement is a means of conserving or improving the overall rock mass properties
from within the rock mass by techniques such as rock bolts and cable bolts. The primary
objective of the reinforcement is to resist the actions induced by stress redistribution within
the rock mass as a result of excavation, thus, reducing shear slip and tensile separation along
rock discontinuities. Rock suspension or “hold” action basically involves the use of rock bolts
and cable bolts to anchor unstable rock blocks that would otherwise fall-out under gravity
loads, thus, addressing local failure mechanisms. Surface retention/support is the application
of a reactive force to the surface of an excavation and includes techniques and devices such as
timber, fill, shotcrete, mesh and steel or concrete sets or liners.
Although tunnel support may involve all three types of support mechanisms concomitantly,
according to Brady and Brown (2004), potential slip on bedding planes is generally the main
problem in the design of a stratified rock mass. The extent is related to the virgin in-situ stress
field and the shape of the excavation. As a result, the primary support mechanism required in
the case of stratified rocks such as the horizontally bedded Hawkesbury Sandstone typically
involves rock reinforcement, and this has certainly be confirmed above.
As a general rule, cases where the span/bed thickness ratio (s/t) is low will be subject to slip
mainly in the haunch area. This may be expressed in the rock mass as the opening of cracks

Figure 6. Primary support functions (after Kaiser et al, 1995).

Figure 7. The effects of slip and separation on excavation peripheral rock (after Brady and Brown,
2004).

5985
Figure 8. Sub-span design approach (after Hutchinson and Diederichs, 1996).

sub-perpendicular to bedding, perhaps coincident with any cross joints in the medium, as illus-
trated in Figure 7a. For a configuration in which the s/t ratio is high (i.e. beds relatively thin
compared with excavation span), the zone of slip may extend further along the span of the
immediate roof. Since the sense of slip on bedding is such as to cause inward displacement
towards the span centreline of beds, the tendency is for isolation of the lower bed, at its
centre, from the one immediately above it. Separation of a roof bed from its uppermost neigh-
bour is highly significant because it implies loss of support of the roof by the overlying beds,
as can be appreciated from Figure 7b. Prior to decoupling of the roof layer, its gravitational
load is carried in part by the more extensive volume of rock in which the layer is embedded.
After detachment of the roof, the bed itself must support its full gravitational load.
The above mechanism is consistent with the failure process described in the previous section
and that of a “voussoir beam analogy”. Based on this mechanism, the primary objective of the
ground support is to provide a reinforcement effect near the abutments to reduce bedding slip
and consequently bedding separation near the mid-span. When possible, rebar bolts are to be
used near the abutments considering their better performance in shear reinforcement.
On the other hand, the tendency of the central zone is to separate from the upper roof beds
once pronounced bedding slip develops near the abutments. As a result, the ground support
within the central zone above crown should focus on a suspension action. In this case the
bolts generally need to be longer than the bolts in the abutment zones bolts in order to achieve
a better anchoring effect above the zone of more significant loss of confinement and roof
movements. As a result, such an effect is often better achieved using cable bolts.
The concept of a suspension type support in central span of the cavern also provides an
added benefit. The cable bolt reinforced central zone not only assists in limiting crown deflec-
tion and dilation but act as an “artificial” abutment for the adjacent spans (Fuller, 1983). For
example, Figure 8 illustrates how that the “line-anchor pattern” provides an “artificial” abut-
ment such that secondary smaller spans develop as marked by dimensions “b” and “h”. This
support strategy is referred to as sub-span design and is often applied in very large mining
openings. It considers the global span to be supported (for the purpose of mining global sta-
bility) although the sub-spans are unsupported. This support strategy is only effective for rock
masses that are primarily dominated by laminations parallel to the excavation face being sup-
ported which is consistent with the horizontally bedded Hawkesbury Sandstone.

5 SUPPORT OPTMIZATION

Based on the assessment of failure discussed in Section 3, the minimum required reinforcement
zone seemed to be approximately a 5 m thick into the tunnel roof. In contrast to previous
designs in Sydney, which used deflection the governing factor that dictates rock beam thickness,
Oliveira and Paramaguru (2016) presented an approach for the design of rock bolts for lamin-
ated rock beams where the main focus is on satisfying the development of the compressive arch

5986
of an equivalent thicker voussoir beam analysis. The rock bolt reinforcement is designed to pro-
vide the necessary capacity to overcome the excess shear stresses in the bedding partings, thus
stitching the thinner beams together into an equivalent thicker rock beam.
The only modification to the proposal by Oliveira and Paramaguru (2016) was the inclusion of
the suspension effect of the longer bolts within the middle zone as targeted in the previous section.
This was done by considering the effect of the “extra-length” of the bolts within the mid-section
of the roof acting as a parabolic pressure as proposed by Diederichs and Kaiser (1999).
The above approach was written in a Mathcad sheet allowing for conveniently verifying
several ground conditions and support spacing to confirm suitability of the proposed support
strategy before more sophisticated numerical modelling.
The effect of ground support was also investigated numerically with discontinuum models
using the Finite Element code RS2 from Rocscience to run a large number of models and dif-
ferent ground conditions and for a few cases using the DEM code UDEC of Itasca. The sup-
port was assessed by including both rock bolt support at the abutment and cable bolt support
in the central zone.
For example, Figure 11 shows an example considering the ground conditions assumed in
the failure mechanism model, with an unfavorable layer of SS-II/III in the tunnel crown.
Although, a tighter bolt spacing would typically be specified for this ground conditions to
control deflections, to assess the potential risk of misclassification during construction, a
wider rock bolt spacing has been modelled. The bolt spacing was adopted at 1.75 m c/c spa-
cing with a 50 mm shotcrete. The rock bolts are 5 m long at the abutments and cable bolts 7.5
m long in the central zone. As a result, some larger roof deflections could be anticipated.
Both multiple heading and full-face excavation sequences have been analyzed but only the
full face sequence discussed. The latter was adopted for comparison with the failure mechan-
ism analysis discussed above.
As observed in Figure 11, the rock support can control the failure mechanism. The pre-
dicted crown deflections vary from approximately 35 mm to about 45 mm within the central
cable bolted zone which results in deflection to span ratios of approximately 0.15%. This
could be considered reasonable for such large spans and comparable to many large concrete
structures spanning such distances, for example in deep excavations where deflections between
0.1% to 0.2% are often observed in stiff fully propped or anchored retaining walls.

Figure 9. Rock reinforcement effect for equivalent rock bolt stitched rock beam (after Oliveira and
Paramaguru, 2016).

Figure 10. Rock reinforcement concept adopted for large span caverns.

5987
Figure 11. Vertical displacements with a supported full-face excavation using 5 m bolts near the abut-
ments and 7.5 m high strength cable bolts within the middle-section both spaced at 1.75 m c/c.

Figure 12. View of 4-lane tunnel in construction (left) and Y-junction cavern (right).

Similar exercise was also carried out for 4-lane tunnels with spans of approximately 20 m.
The main difference between the 4-lane tunnels and the caverns was that the former did not
require the use of the longer cables in the middle zone with only 5 m long rock bolts used
throughout the span but still 1 m shorter than the precedent design.

6 CONSTRUCTION

As previously discussed, these large tunnels are currently in construction. However, the avail-
able monitoring data from 3D prism convergence (survey), multipoint extensometers and
instrumented rock bolts have generally confirmed the predicted tunnel performance using the
proposed approach. Most of the large span roof deflections have been observed to be between
10 mm to 50 mm, depending on several factors such as rock quality, excavation sequence, in-
situ stress magnitude and orientation etc.

7 CONCLUSIONS

This paper presented a discussion on the design of large span tunnels in Sydney Australia.
Some precedent design was presented and demonstrated that, although very valuable, past
projects are not the only single source of design.
Design efforts that focus on first principles and the basic objectives of rock reinforcement
have overcome the challenges of these designs, satisfied codes and standards requirements but
at the same time provided savings with respect to ground support. For example, tunnels spans
of approximately 20 m were demonstrated to be satisfactory with 5 m long cables in contrast
to a precedent design value of 6 m long bolts. Large span caverns also have the maximum bolt

5988
length of 7.5 m in its mid-span with shorter bolts at the abutments compared to an precedent
design of 9 m long bolts.
The key to the design involved understanding the key failure mechanisms that needs to be
addressed, its relationship with the different actions of rock bolting, i.e. suspension/anchorage
and/or rock reinforcement and what could be considered acceptable for design.

REFERENCES

Bertuzzi, R. and Pells, P.J.N. (2002). Design of rock bolt and shotcrete support of tunnel roofs in Sydney
sandstone. Australian Geomechanics, 37(3).
Brady, B. H. G., and Brown, E. T. 2004. Rock Mechanics for Underground Mining. Dordrecht: Kluwer
Academic Publishers.
Diederichs, M S and Kaiser P. K. (1999): Stability Guidelines for Excavations in Laminated Ground -
The Voussoir Analogue Revisited, Int. J. Rock Mech. & Min. Sci.; 36, pp 97–118.
Fuller, P.G. (1983) Cable support in mining: A keynote lecture. Rock Bolting, Rotterdam: A.A. Balkem,
511–522.
Hutchinson, D.J. and Diederichs, M.S. (1996). Cablebolting in Underground Mines. Bitech Publishers
Ltd., Canada.
Kaiser, P.K., McCreath, D.R. and Tannant, D.D. (1995). Rockburst support handbook. Geomchanics
Research Centre, Sudbury.
Oliveira D. and Paramaguru, L. (2016). Laminated rock beam design for tunnel support. Australian
Geomechanics Vol 51(4),pp.1–17, December.
Pells, P.J.N., Mostyn, G., Walker, B.F. (1998). Foundations on sandstone and shale in the Sydney
region. Australian Geomechanics; Vol 33(3):17–29.

5989

You might also like