Tugun Bypass Tunnel Using Top-Down Cut-and-Cover Method

J Hsi1, S Lambert2 and M Thomas3

ABSTRACT
The Tugun Bypass Tunnel in Gold Coast, Australia was constructed using diaphragm walls with the top-down cut-and-cover method to allow simultaneous construction of an airport runway extension above the tunnel, whilst excavation of the tunnel continued underneath. The tunnel was built in an environment of high groundwater table and deep deposits of alluvial and estuarine soils with the toes of the walls founded in soil deposits. There was a potential risk for differential settlements to occur between the diaphragm wall panels, caused by the runway fill placed over the tunnel roof during excavation. Dewatering within the diaphragm walls was required to facilitate the construction of the tunnel. The tunnel was also built in an area where environmental considerations were of great importance. Three-dimensional numerical modelling was undertaken to predict the differential settlements of the tunnel with considerations of varying subsurface profile, staged excavation and dewatering, nonuniform surface loading and complex soil-structure interaction. Field instrumentation and monitoring was implemented to confirm numerical predictions.

South Wales, Australia. The highway deviates off the existing Pacific Motorway at Stewart Road, traverses hilly terrain through Tugun Hills and floodplains adjacent to the Gold Coast Airport, and merges with Pacific Highway at Kennedy Drive. The key feature of the project is the tunnel of 334 m in length (Ch5588 to Ch5922.4), with approach ramps to the north and south of the tunnel. The project also involves cuttings of up to 30 m in depth, five bridges, four soil nail walls and a large quantity of earthworks. The bypass caters for potential expansion to a six lane motorway subject to future traffic volume. Figure 1 shows the project route plan.
CH100 CH100

INTRODUCTION
The Tugun Bypass is a new four-lane motorway of approximately 7 km in length, connecting south-east Queensland and north-east New South Wales, Australia. One of the main features of the project is a tunnel of about 334 m in length, constructed near one end of the Gold Coast Airport runway. The tunnel was constructed using the top-down cut-and-cover construction method with diaphragm walls installed to support temporary excavation and form permanent walls of the tunnel. As the tunnel was constructed near one end of the runway, there was a height restriction on the construction plant and equipment. Twin low headroom cutters and a hydraulic grab were selected for the construction of the diaphragm walls. These walls were constructed in alluvial and estuarine deposits comprising sands and clays with the groundwater table close to the surface. During excavation of the tunnel, the airport runway was to be extended above the tunnel, which involved filling over the tunnel roof. The loads acting on the roof were carried by the diaphragm walls, which were located entirely within the alluvial and estuarine deposits. Some settlements would occur, resulting from the additional loading of the runway extension and loss of roof slab bearing and wall skin friction due to excavation. To maintain the wall and roof structural integrity, differential settlements in the longitudinal and transverse direction of the tunnel needed to be minimised. This paper also discusses other key features of the tunnel, including environmental management in sensitive areas, dewatering and recharging of the groundwater during excavation, diaphragm wall construction methods, waterproofing of the tunnel, challenges of working under the obstacle limitation surface (OLS), and cathodic protection of the reinforcement.

CH5588
CH5922.4 CH5922.4

CH7200 CH7200

FIG 1 - Project route plan.

PROJECT OVERVIEW
The 7 km long Tugun Bypass is a four-lane motorway connecting Currumbin in Queensland to Tweed Heads in New
1. 2. 3. Chief Technical Principal, SMEC Australia Pty Ltd, Level 6, 76 Berry Street, North Sydney NSW 2060. Email: jeff.hsi@smec.com.au Project Manager, Abigroup Ltd, 924 Pacific Highway, Gordon NSW 2072. Email: steve.lambert@pacificlinkalliance.com Senior Tunnel Engineer, SMEC Australia Pty Ltd, Level 2, 60 Leichhardt Street, Spring Hill Qld 4000. Email mike.thomas@smec.com.au

The Tugun Bypass project was awarded to the PacificLink Alliance (PLA) in January 2006 under an alliance style contract following a competitive tendering process. The alliance team, consisting of Department of Main Roads Queensland, Abigroup Contractors and SMEC Australia, was responsible for the design and construction of the project. A suballiance, consisting of Piling Contractors Bauer Joint Venture (PCBJV), was engaged to construct the tunnel diaphragm walls. The estimated value of the project was approximately A$500 million. The project met the deadline of November 2006 for the surface handover to the Gold Coast Airport Authority for construction of the runway extension. The remaining construction works continued, targeting the scheduled contract completion date of December 2008. Following commissioning of the road, there is a ten year maintenance period as part of the alliance contract.

GEOLOGY AND GEOTECHNICAL CHARACTERISTICS Geology

The Tugun Bypass alignment is characterised by two geological settings. In the northern section, the alignment traverses hilly terrain consisting of Neranleigh Fernvale Beds of the Beenleigh

13th Australian Tunnelling Conference

Melbourne, VIC, 4 - 7 May 2008

J HSI, S LAMBERT and M THOMAS

block comprising lithic sandstone (referred to as greywacke), slate, metasiltstone (argillite) and chert. The rocks developed in these zones have been subjected to varying degrees of deformation and typically identified by coherent strata being discontinued by a tectonic mlange. Highly sheared material up to 10 m wide, which contains varying size blocks of country rock (eg chert, greenstone, siltstone and sandstone) bounding the mlange zone, are typical. Groundwater is not usually encountered in the northern section. The tunnel is situated in the southern flood plain, which is subject to periodical flooding. The geology in the southern section comprises Neranleigh Fernvale Beds overlain by Cenozoic estuarine and coastal deposits. These deposits are up to 35 m in thickness, comprising river gravels, sand and clay and flood plain and tidal delta muds and silts. At the tunnel location, the subsurface horizons consist of dune sands, Coffee Rock (local term given to cemented silty sands), estuarine interbedded clays and sands and residual soils derived from the weathered bedrock. Groundwater is slightly saline due to the close proximity to the ocean. The water table is influenced by both tidal movements and rainfall events recharging Cobaki Broadwater. Due to low-lying ground surfaces ranging from RL 0.5 m to RL 4.0 m (AHD), potential exists for acid sulfate soils.

Dune sands a sequence of generally loose to very loose

sands of up to about 8 to 10 m in thickness, fine- to mediumgrained sands.

Coffee rock a sequence of medium dense to very dense

cemented silty sands of about 7 to 10 m in thickness with occasional loose consistency.

Estuarine a sequence of about 15 m thickness comprising

shell fragments, sand and silty sand, clay and sandy clay, silt and clayey silt, clayey silty sand and gravels. Relative density varied from very loose to dense, and consistency varied from firm to very stiff.

Residual soil comprising clay and silty clay with some

sands, and with residual fragments of extremely weathered and extremely low strength interbedded argillite and greywacke of the Neranleigh Fernvale Beds. The thickness ranged between about 1 m and 6 m.

Bedrock comprising extremely weathered to moderately

weathered and extremely low to low strength interbedded argillite and greywacke, having an irregular contact with the overlain residual material at a depth of approximately 30 to 35 m. The subsurface profile based on the boreholes along the centre line of the tunnel under the runway extension is presented in Figure 3.

Subsurface profile
As the subsurface conditions varied spatially along the length and width of the tunnel, extensive site investigations using boreholes (BH) and piezocones (CPTU) were undertaken at the wall and barrette locations. Within the footprint of the runway extension, the investigations were done at a spacing of approximately 20 m intervals. The plan of the site investigation is shown in Figure 2. Based on the probe hole information, the subsurface was divided into discrete soil units, classified according to material type and consistency or density. There was Coffee Rock (CR) found in the tunnel areas. Coffee Rock is a local term used to describe a layer of cemented silty sand having dark brown coffee-like colour. Contrary to its name, this material has soil characteristics, and its relative densities are typically medium dense or better, although loose Coffee Rocks were also encountered occasionally. Geotechnical stratigraphy at the tunnel site is summarised as follows (top down):

Geotechnical parameters
The geotechnical parameters for each of the units were determined from interpretation of the field and laboratory test results, and based on local experience. The key parameters included stiffness and strength parameters. As undisturbed samples for sandy soils were nearly impossible to obtain, their parameters had to be estimated based on published values for similar conditions. Oedometer tests on clay samples were undertaken to derive compressibility of the clay. Coffee Rock was assumed to be permeable despite its cementation nature. Also, as Coffee Rock had higher silt content compared to dune sand, it was assumed to have lower strength compared to sands of similar relative densities. Based on the interpretation of the geotechnical data, the geotechnical parameters assumed at approximately Ch5750 are shown in Table 1. The table also shows the indicative reduced level (RL) at the top of each of the layers, and the corresponding range of thickness. The ground surface level was approximately at RL 0.5 m and the groundwater table was at surface.

Topsoil thin skinned (<1 m) comprising peaty sandy

organic topsoil, having loose consistency. The ground surface was marshy and generally untrafficable.

FIG 2 - Site investigation plan.

Melbourne, VIC, 4 - 7 May 2008

13th Australian Tunnelling Conference

TUGUN BYPASS TUNNEL USING TOP-DOWN CUT-AND-COVER METHOD

FIG 3 - Subsurface profile.

TABLE 1
Geotechnical parameters.
Average RL at top (m) 0.5 -4.0 -11.2 -13.5 -17.5 -21.1 -23.0 -28.6 -30.8 Thickness range (m) 3.3 - 5.5 4.8 - 11.0 1.5 - 3.0 1.2 - 5.8 0.0 - 6.5 0.5 - 3.3 3.2 - 7.2 1.3 - 3.0 Material type Sand CR CR Sand Clay Sand Clay Clay Bedrock Consistency/ density Very loose Medium dense Dense Loose Stiff Loose Firm Very stiff (kN/m3) 18 19 20 18 18 18 18 19 sat

' ( 30 32 34 32 28 32 24 29 -

' () 0 2 4 2 0 2 0 0 -

k (m/day) 1.0 0.1 0.1 1.0 1 10 1.0 1 10- 4 1 10- 4 -4

E50ref and Eoedref (MPa) 10 50 80 30 10 30 7 25 ref

Eurref (MPa) 30 150 240 90 30 90 21 75 -

Note: sat is saturated unit weight; ' is drained friction angle; ' is drained dilatation angle; k is saturated permeability; E50 is secant Youngs modulus at a reference pressure of 100 kPa; Eoedref is tangent Youngs modulus for primary odometer loading at a reference pressure of 100 kPa; Eurref is unloading/reloading Youngs modulus at a reference pressure of 100 kPa; cohesion c= 0 kPa for all soil types; Poissons ratio = 0.3 and ur = 0.2 (unloading/reloading) for all soil types; power m used in hardening soil model for stress level dependency is 1.0 for clay and 0.5 for Coffee Rock and sand.

ISSUES AND CONSTRAINTS

Construction of a tunnel in soft ground at shallow depths is conventionally undertaken using the cut-and-cover method. However, to allow for construction of the runway extension that occurred concurrently with the tunnel excavation, the top-down construction method was employed. Diaphragm walls and cast in situ tunnel roof slabs were chosen to facilitate the construction requirements and time constraints. Figure 2 shows the footprint of the runway extension oblique to the tunnel alignment. Following the handover of ground surface, up to 2 to 3 m of fill for the airport runway extension was placed above the tunnel roof. Loads acting over the entire width of the roof slabs were transferred directly to the diaphragm walls and the barrettes. The site investigations revealed presence of estuarine deposits consisting of loose materials below the toe of the walls. Therefore, there was a potential for the tunnel to settle during excavation. One of the critical issues was the differential settlements between the walls and the central barrettes, and along the walls. These differential settlements could potentially induce significant stresses in the roof structures and in the walls. Other issues in relation to the tunnel construction are listed below:

This required all construction activities to be undertaken within a headroom of as low as 8 m. Use of cranes or heavy-lifting equipment was only allowed outside the airport operating hours.

High groundwater level due to its close proximity to the sea

and Cobaki Broadwater. The groundwater was practically at the ground surface level. A reliable dewatering system was essential during excavation. were enforced such that drawdown of the groundwater table outside the diaphragm walls was minimised. All acidic sulfate soils excavated from the tunnel had to be dried and neutralised with lime prior to placement as fill in embankments. The existing ecological conditions also needed to be enhanced.

Environmental requirements strict environmental controls

CONSTRUCTION METHODS
Suitable construction methods were chosen to address the issues and constraints mentioned above. In order to adhere to the OLS requirements, special low headroom hydraulic grab (Leibherr HS852HD) and 2.8 m wide trench cutter (CBC25) were used. The guide walls were built first followed by construction of the 6 m wide primary panels (steps one to five of Figure 4) and 2.8 m

Obstacle limitation surface (OLS) applied at both ends of

the runway to provide safe airspace for approaching aircraft.

13th Australian Tunnelling Conference

Melbourne, VIC, 4 - 7 May 2008

J HSI, S LAMBERT and M THOMAS

FIG 4 - Diaphragm wall construction sequence (courtesy of Piling Contractors Bauer Joint Venture).

wide secondary panels (steps six to eight of Figure 4). The open trench was supported by bentonite slurry, when the cutter undertook full excavation (steps two to three and six). A steel reinforcement cage was lowered when the panel was excavated to full depth (steps four and seven). Concreting of the panels was then achieved by the tremie method (steps five and eight). Figure 4 presents the construction sequence of the diaphragm wall. Following completion of the diaphragm walls and barrettes, dewatering and excavation commenced inside the walls. Excavation was initially undertaken to depths of up to about RL -2 m to allow construction of the roof slab. A watertight membrane was installed as part of the waterproofing system. When the roof slab was completed, it was backfilled and the site was cleared for handover to the Gold Coast Airport. These activities commenced in April 2006 after environmental approvals were granted and were completed by November 2006, which was the scheduled date of the site surface handover. Excavation below the runway extension and construction of the tunnel continued thereafter.

slabs were cast in situ structures connected to the diaphragm walls by reinforcement couplers in a typical moment connection. The clear height was 6.1 m with minimum vehicular headroom of 5.3 m within the carriageway envelope. A series of three niches, where the roof slab was slightly elevated, were constructed in the roof for mechanical jet fan provision. Niche dimensions were in plan view 23 m long across the full width of the tunnel and were evenly spaced along the length of the tunnel. The thickness of the overburden including the runway pavement was up to 4 m and decreased to 1 m at the tunnel portals. The base slab was 1 m thick with a founding level ranging from RL -5.5 m to RL -9.5 m. The base slab had all drainage provision cast monolithically within the slab to mitigate potential for pipe breakages caused by differential movement and buoyancy forces. Figure 5 shows the typical cross-section of the tunnel. Cathodic protection provision, sheet and joint waterproofing, mechanical/electrical service provision and monitoring instrumentation were also cast into the tunnel structure.

Approach ramps
The north and south approach ramps were both constructed by the bottom-up construction method. The south ramp was 281 m long and 24 m wide at the portal tapering out to 44 m at the top of the ramp to accommodate a slip lane. The north ramp was 271 m long and 28 m wide. Diaphragm walls were used to form part of the permanent ramp structures within 32 m from the portal, followed by temporary sheetpiles to allow construction of the ramps in the narrow corridor. The extended diaphragm walls were 27 m deep and made up one wall of the deep sumps, which were located at each portal. Sump dimensions were 31 m long (longitudinal) and 16 m wide (transverse) with an average depth of 5.5 m below the base slab and were cast in situ elements. One side of the sump was coupler connected to the extended diaphragm walls. The walls and ramp slabs were cast in situ elements with provision for cathodic protection, waterproofing, monolithic drainage, mechanical/electrical services and tension screwpiles in the sump and ramp slabs.

STRUCTURAL DETAILS Tunnel box

The tunnel structure consisted of diaphragm walls forming the outer walls and barrettes along the centre line of the tunnel. The diaphragm walls were 1 m in thickness and extended from the Northern Portal (Ch5588) to the Southern Portal (Ch5922.4). The walls were installed to the depth of RL -17 m from the top of the guide wall at RL 2 m. The internal width between the diaphragm walls ranged from about 25.7 m at the northern portal to 28 m at the southern portal. Barrettes were 0.8 m thick and 2.8 m wide (longitudinal) with a clear spacing of 2.8 m throughout the central axis of the tunnel, extending to RL -17 m in depth. These structures had a 100 year design life, using N-grade reinforcing steels and 50 MPa high strength concrete. There were no mechanical joints at the interface of the primary and secondary panels in the longitudinal direction. The roof and base

Melbourne, VIC, 4 - 7 May 2008

13th Australian Tunnelling Conference

TUGUN BYPASS TUNNEL USING TOP-DOWN CUT-AND-COVER METHOD

Screw piles
Tension piles were required to prevent flotation of the approach ramps. The piles were designed for loads from the one in 100 year flood and a design life of 100 years. Maximum upward movement under long-term load was limited to 25 mm to ensure that the interface stresses between the ramp and tunnel portal did not exceed allowable values. Steel screw piles were selected due to the tight site constraints, the least noise impact on neighbouring properties and the significant cost advantages resulting from being a very efficient pile in resisting tension. Piling could be carried out day and night with simple equipment (excavator rig and hydraulic motor), which provided flexibility to the construction program. Site trials were performed to determine the most appropriate parameters for the pile. The typical configuration adopted was a shaft of diameter 219 mm and thickness 8.2 mm, fitted with a helix of a diameter of 600 mm and a thickness of 32 mm. In weaker ground the helix was increased to 700 mm diameter and preboring was required in areas where dense or very dense sand were encountered in order to overcome the installation limits. For longer piles where preboring was not possible the shaft thickness was increased to 12.7 mm for increased installation torque capacity.

A sacrificial thickness of 4 mm was incorporated to allow for corrosion over the 100 year design life. The working load was in the range of 300 to 350 kN and typical pile lengths were 9.0 to 13.5 m, with rows of eight to ten piles placed at spacings varying from 2.4 to 4.0 m. A typical cross section of screw pile arrangement is shown in Figure 6. Installation torques were used to control pile founding depths based on results from preliminary test piles where a correlation was derived between torque and SPT values in various ground conditions. Corrosion monitoring elements were installed at 12 locations throughout each ramp, set within 0.5 to 1 m of the piles and at varying depths of up to 7 m. At three locations additional sacrificial piles were installed each side of the ramp with a detailed, accurate recording of pile section thickness. These piles could later be extracted for measurement as part of the overall corrosion monitoring system. Bonding bars connected the reinforcement at the top of each pile (which was isolated from the ramp slab reinforcement) to allow for future use of cathodic protection if required.

Waterproofing
The approach ramps and tunnel were made watertight by a fully bonded membrane system applied to the permanent structure

Roof Slab

Diaphragm Wall

Waterproof Membrane

Waterproof Membrane

Base Slab Barrette

FIG 5 - Typical tunnel cross-section.

Base Slab

Screw Pile

FIG 6 - Screw piles under approach ramp.

13th Australian Tunnelling Conference

Melbourne, VIC, 4 - 7 May 2008

J HSI, S LAMBERT and M THOMAS

externally (see Figure 5). The diaphragm walls did not receive an external membrane, but rather a watertight joint between the primary and secondary panels was achieved through the joint treatment (ie tongue and groove) from the CBC25 low headroom diaphragm wall cutter, as shown in Figure 7.
cutter

Cathodic protection
The tunnel diaphragm walls and barrettes were protected against corrosion by an impressed current cathodic protection (ICCP) system using anodes connected to a dc power source. The system adopted a total of 58 vertical anode wells extending to 30 m depth. The Wenner Method of measuring soil resistivities was adopted to gather the necessary design parameters of depth, thickness and resistivity of soil layers. The positive dc output terminal was connected via cables to the anode array while the negative output was connected to the diaphragm wall and barrette reinforcing steel bars. For typical arrangement of the ICCP system refer to Figure 8. A corrosion monitoring system was implemented for the approach ramps and tunnel slabs utilising a Moncor Corrosion Monitoring unit. The system was capable of providing information to assess the level of corrosion of steel reinforcement and early detection of chloride contamination. The monitoring units were imbedded within the concrete at representative locations in the structure and took measurements of the relative degree of chloride ingress into the concrete, the corrosion rate and the corrosion activity potential of the steel reinforcement. The levels of relative chloride ingress and corrosion rate were measured at a junction box using a specialised external instrument. A monitoring system was also installed to the approach ramp screwed piles, with provision made for the future connection of the screw piles to an ICCP system. The monitoring system located a corrosion probe, zinc reference, silver/silver chloride reference and titanium reference electrode at depth beneath the permanent structure. Electrical continuity was maintained via the steel reinforcement within the concrete structures with a negative connection welded to the steel screw piles. The probes, electrodes and negative connections terminated at junction boxes for monitoring using special equipment for corrosion rate monitoring and reference electrodes potential.

secondary panel primary panel primary panel H - beam to be removed during concreting

FIG 7 - Joints between diaphragm wall panels (courtesy of Piling Contractors Bauer Joint Venture).

The 1.5 mm thick membrane Bituthene 3000, manufactured by Grace Construction Products, was cold applied in two layers. The self-adhesive membrane, with 100 micron cross-laminated HDPE film and rubber bitumen compound, was applied to 75 mm thick blinding concrete beneath slabs, or directly to external face of approach ramp walls. The membrane in the wall application was protected by protection board placed prior to the backfilling operations. All joints within the structure received a hydrophilic waterstop, with a volumetric increase of greater than 200 per cent when in contact with water. In addition, an injection tube system capable of reuse and through which both grout and resin injection could take place was located adjacent to the waterstop.

DEWATERING SYSTEM
The dewatering system for the tunnel works consisted of dewatering wells, a chemical treatment system, settling ponds and recharge wells. The system was required to allow construction of the tunnel, which at its deepest was more than 10 m below sea level and at its closest point was only 60 m from the Cobaki estuary system.

FIG 8 - Cathodic protection arrangement.

Melbourne, VIC, 4 - 7 May 2008

13th Australian Tunnelling Conference

TUGUN BYPASS TUNNEL USING TOP-DOWN CUT-AND-COVER METHOD

The dewatering system was installed in stages to suit the complex nature of the tunnel construction. The initial dewatering wells were installed to permit excavation for roof construction. The dewatering pipe work for the main tunnel excavation was installed under the roof slab such that in the next stage of works the dewatering pipe work would be suspended from the roof slab. This enabled dewatering of the main tunnel zone ahead of the excavation works and allowed the excavation works to progress without disruption to the dewatering system. A total of 21 dewatering wells were installed with the deepest wells installed to RL -25 m, approximately 10 m below the deepest tunnel excavation level. The extraction wells consisted of an outer screened casing, coarse sand filter and an inner extraction pipe. Air pumps at the base of each well lifted the groundwater to the surface where it was piped to one of the two treatment plants at either end of the tunnel. Electric and diesel compressors at each of the treatment plants provided the compressed air to run the pumps. Air pumps were chosen because of their simplicity. They had no moving parts and as such were very reliable and essentially maintenance free. The extracted water was high in iron (40 to 60 mg/L) and if left untreated would quickly clog the recharge wells with iron precipitates. The pH and iron levels were also unsuitable for direct surface discharge, so that treatment was essential. In the treatment plant the pH of the groundwater was raised to 8.0 to 9.0 by injection of a caustic soda (NaOH) solution. At the elevated pH the iron began to precipitate out of the solution. The reaction was relatively fast, so after a short contact time the pH was brought back down to 6.5 to 7.5 using hydrochloric acid (HCl). This halted the process of iron precipitation and also reacted with the caustic soda to form common salt (NaCl). At this stage the iron had come out of the solution and was a suspended solid. The processed groundwater was then pumped to a series of settlement ponds where the iron floc was allowed to settle out. Additional treatment could be conducted in the ponds to improve turbidity or to adjust the pH to meet the discharge criteria. The treated water was then pumped to the recharge system. The recharge system consisted of a storage tank, used to limit and maintain a constant head to the recharge wells, and a bank of recharge wells connected by a system of pipe work, both air and water. There were two separate recharge systems for each end of the tunnel. The treated water was gravity fed from the header tank to the recharge wells. These consisted of a screened casing surrounded by a gravel pack installed to a depth of around 20 m. The top of the well used unscreened casing and a bentonite plug to allow the well to be pressurised. The pressure was limited to 1 - 2 m of head as the natural ground consisted mainly of remnant dune sands and under high pressures the sand tended to liquefy and could blow out at the surface. Recharge flows were generally in the range of 1 - 1.5 L/s per well. Excess water from the recharge system was discharged as surface water at designated discharge points. Although most of the iron was removed in the treatment process and in the settling ponds the recharge wells needed to be back flushed at regular intervals to remove iron build-up on the screens and in the gravel packs. This was done using compressed air, which was fed to the bottom of the well by an internal air line. The iron-rich water, which was back flushed to the surface, was then returned to the dewatering system. The air used to back flush the wells was supplied from the same compressors that powered the air pumps for the dewatering system. As the tunnel works progressed and sections became fully sealed the dewatering wells were decommissioned by cutting the discharge pipe at the top of the tunnel base slab, the air pump was extracted, the conduit sealed with an expandable bladder and the hole grouted back to slab level.

GEOTECHNICAL ANALYSIS Design considerations

Geotechnical design of the tunnel was required to address the following three key issues:

Excavation support during construction the diaphragm wall

structures were designed to ensure stability of the excavation. Issues included structural design of the walls, base heave, hydraulic uplift, piping and liquefaction.

Structural integrity due to settlement of the tunnel during

construction the tunnel was subjected to loading from airport runway fill, which resulted in settlements. The influences of differential settlements on structural capacity were assessed.

Long-term stability and serviceability of the tunnel

buoyancy of the tunnel when the groundwater table was close to the surface or the flood level.

Two-dimensional numerical modelling

Design of the tunnel was initially undertaken using the finite element software PLAXIS (Version 8.4) at selected sections. This numerical package was used to analyse two-dimensional plane-strain conditions involving complex soil-structure interaction for the design of the structural members. Structural beam elements were used to simulate the diaphragm walls. Global factor of safety during each of the construction stages was calculated based on the c- reduction method to ensure the minimum FoS was achieved. The software allowed modelling of construction sequence, changing groundwater levels and varying subsurface conditions across the width of the tunnel.

Three-dimensional numerical modelling

A three-dimensional numerical modelling package, PLAXIS 3D Foundation (Version 1.6), was employed to predict the settlements of the tunnel caused by runway fill loading and excavation. Due to the limitation of the program, settlement analyses were undertaken in sections, each of approximately 40 60 m in length. The major advantages of the three- dimensional modelling were as follows:

Ability to model the physical dimensions of the wall and

barrette structures. This improved the accuracy of settlement prediction as it accounted for longitudinal stiffness of the tunnel, which assisted in load redistribution and toe resistance of the structures.

Ability to simulate three-dimensional load distribution where

the runway fill was placed oblique to the longitudinal axis of the tunnel. on probe holes at discrete locations.

Ability to model three-dimensional subsurface profile based Ability to simulate dewatering within the tunnel excavation
area. The hardening soil (HS) model was considered most appropriate to simulate soil behaviour in an excavation. The HS model took into account unloading and reloading behaviour and irreversible plastic strains of soil. The HS stiffness parameters were defined with respect to a reference pressure of 100 kPa. The key parameters included E50ref, Eoedref and Eurref as shown in Table 1. The published data indicated the ratio of Eoedref to E50ref is about 0.7 to 1.4 and the ratio of Eurref to E50ref varied from two to four. The analysis adopted E50ref = Eoedref and Eurref = 3E50ref. Presented here is a 41.2 m long section of the tunnel between Ch5728.8 and Ch5770. This section of the tunnel was at the deepest location of the tunnel, beneath the thickest layer of the

13th Australian Tunnelling Conference

Melbourne, VIC, 4 - 7 May 2008

J HSI, S LAMBERT and M THOMAS

runway fill and underlain by sloping bedrock level and changing clay thickness. A jet fan niche of approximately 12 m long (longitudinal) also lay within the centre of this section, which was also incorporated in the three-dimensional model. Within this chainage range, there were seven boreholes. Due to the capacity of the program, four representative boreholes, which were evenly distributed spatially, were selected for the analysis. The assumed subsurface profiles are shown in Table 2. TABLE 2
Subsurface profiles.
Borehole Location Chainage Unit 1 2 3 4 5 6 7 8 9 0.5 -4.8 -10.8 -13.8 -17.8 -24.3 -27.3 -33.3 -36.3 #1 LHS 5730 #2 RHS 5737 0.5 -3.4 -14.4 -15.9 -17.1 -21.8 -25.1 -30.8 -32.0 #3 LHS 5757 0.5 -5.0 -9.8 -12.6 -17.5 -20.8 -21.8 -25.0 -27.4 #4 Centre 5768 0.5 -2.8 -9.7 -11.7 -17.5 -17.5 -18.0 -25.2 -27.4 Sand (VL) CR (MD) CR (D) Sand (L) Clay (St) Sand (L) Clay (F) Clay (VSt) Bedrock Soil type (density/ consistency)

4. 5. 6. 7. 8. 9.

dewatering and excavation to underside of the roof slab; installation of the roof slab (and jet fan niche), and backfill to existing ground surface; placement of runway fill to design heights (simulated as pressures) with 10 kPa live load above the runway; staged dewatering and excavation within the diaphragm walls to underside of the base slab; casting of the base slab and completion of the tunnel structure; and return of the groundwater table to the ground surface and removal of 10 kPa surface loads.

RL at top of each layer

The settlement-induced impact was assessed for the above stage seven, which was considered most critical with maximum excavation under full runway loading. The assumed levels within the modelled chainage range are summarised in Table 3.

Results of analysis
The deformed mesh of the three-dimensional finite element analysis under the full runway loading and at the final stage of the excavation is shown in Figure 9. The predicted settlement profiles (Class A prediction) at the top of the roof slab along the diaphragm walls and barrettes prior to casting of the base slab are presented in Figure 10. The predicted settlement (Class A) of the tunnel during excavation was about 45 mm on the left-hand side (LHS), 43 mm on the right-hand side (RHS) and 35 mm along the central barrettes. The maximum differential settlement was predicted to be 12 mm between the walls and the barrettes. To allow for uncertainties, the tunnel was designed for a maximum differential settlement of 25 mm. The structural analysis showed that the longitudinal in-plane stiffness of the tunnel would smooth out differential settlements along the tunnel alignment, with the presence of the jet fan niche and variability of the subsurface conditions.

Note: LHS is left-hand side of tunnel facing increasing chainage direction; RHS is right-hand side of tunnel; centre is centre line of tunnel.

Assumptions of analysis
The construction sequence was considered in the analysis to simulate the load transfer from the runway fill to the diaphragm walls and barrettes. The assumed construction sequence is described below: 1. application of loads exerted on the virgin ground from the working platform built to RL 2 m (for construction of the guide walls) and construction load of 10 kPa; installation of diaphragm walls and barrettes to RL -17 m; removal of the working platform and application of 10 kPa construction load on the ground surface;

Field performance
The performance of tunnel during construction was assessed based on the field monitoring results. This was a means to confirm that the structural integrity of the diaphragm walls and barrettes were not adversely affected by differential settlements.

2. 3.

TABLE 3
Assumed geometry during construction.
Chainage range Feature Natural ground level Top of roof slab Bottom of roof slab Initial excavation Initial dewatering Intermediate excavation Intermediate dewatering Top of base slab Bottom of base slab Final excavation Final dewatering Toe of diaphragm wall 0.5 -0.8 -1.8 -2.8 -3.8 -6.0 -7.0 -8.4 -9.4 -9.7 -11.7 -17.0 0.5 -0.25 -1.25 -2.8 -3.8 -6.0 -7.0 -8.4 -9.4 -9.7 -11.7 -17.0 5728.8 to 5737.6 5737.6 to 5743.6 5743.6 to 5755.2 RL (m) 0.5 +0.4 -0.6 -2.8 -3.8 -6.0 -7.0 -8.4 -9.4 -9.7 -11.7 -17.0 0.5 -0.25 -1.25 -2.8 -3.8 -6.0 -7.0 -8.4 -9.4 -9.7 -11.7 -17.0 0.5 -0.8 -1.8 -2.8 -3.8 -6.0 -7.0 -8.4 -9.4 -9.7 -11.7 -17.0 5755.2 to 5761.2 5761.2 to 5770.0

Melbourne, VIC, 4 - 7 May 2008

13th Australian Tunnelling Conference

TUGUN BYPASS TUNNEL USING TOP-DOWN CUT-AND-COVER METHOD

FIG 9 - Deformed three-dimensional finite element mesh.

Three instrumentation arrays were set up at Ch5655, Ch5718 and Ch5770, corresponding to locations of the runway fill (see Figure 11). Each array consisted of three settlement plates placed above the LHS and RHS diaphragm walls and the central barrettes (see Figure 12). These were installed prior to runway fill placement and excavation of the tunnel in order to capture all constructioninduced movements. In addition to the settlement plates, survey targets were also installed at inner walls to the tunnel to record tunnel movement during excavation. This information was calibrated against the settlement plate measurements as the initial tunnel movement record was not available. Figure 13 shows a summary of construction activities, recorded settlements and the predicted settlements at diaphragm wall and barrette locations at Ch5718. The settlement prediction adopted here is the result of analysis between Ch5728.8 and Ch5770. Monitoring commenced at the beginning of November 2006. Excavation of the tunnel commenced in mid December

Distance along Centre Line (m)

0 -30 -5 -10 -15 -20 -25 -30 -35 -40 -45

Settlement (mm)

-35

-40

-45

-50 CH5728.8

Central Barrettes RHS Diaphragm Wall LHS Diaphragm Wall

CH5770

-55

FIG 10 - Predicted settlement profiles at top of roof (Class A prediction).

FIG 11 - Plan of instrumentation arrays.

13th Australian Tunnelling Conference

Melbourne, VIC, 4 - 7 May 2008

J HSI, S LAMBERT and M THOMAS

FIG 12 - Typical instrumentation section.

2006 from the Northern Portal at Ch5588. The excavation process reached Ch5718 in early January. Placement of runway fill above CH5718 followed in mid January, which had resulted in visible settlements of the tunnel. The settlements appeared to have ceased after the excavation reached final depth in mid February. The monitoring data showed that the field performance of the tunnel was consistent with the predictions obtained from the PLAXIS 3D foundation modelling. Maximum differential settlements between the barrettes and the diaphragm walls were less than 25 mm at all stages of construction.

licensed disposal facility. The remedial works minimised the potential migration of groundwater pollution from the area and groundwater monitoring was undertaken quarterly to confirm that ongoing contaminant migration did not occur. A total of 12 500 m2 of concrete paving was dug up, crushed and reused on site within culverts as an anti-erosion measure and as general fill throughout the site. Approximately 200 m3 of construction waste was taken off site every month and sent to a recycling depot where 80 per cent of it was recycled. Up to one million litres of water a day was required during hot summer months for dust control on site. Where possible this water was sourced from sediment basins where site run-off was collected. All mulched vegetation was reused on site for the landscaping and revegetation and also for the stabilisation of exposed topsoil and substrate. Vegetation clearing along the alignment was undertaken in a controlled and methodical manner, with very few breaches of the delineated clearing limit. A combination of well signposted clearing limits and ongoing education of staff ensured that all were aware of the environmental significance of the surrounding vegetation and fauna habitat. Environmental management measures were implemented prior to any clearing works and these included the provision of a professional fauna handler to deal with any fauna found during the clearing. Treatment of the groundwater associated with the tunnel works is described above in the Dewatering System section.

CONCLUSIONS
The Tugun Bypass tunnel was constructed under many strict constraints, including the obstacle limitation surface (OLS), thick alluvial and estuarine soil deposits, groundwater table at shallow depths, sensitive environment, early handover of site surface for runway extension, etc. Diaphragm walls with the top-down construction techniques were adopted for the construction of the tunnel structure. Detailed geotechnical investigations and numerical modelling were undertaken for design optimisation and risk minimisation. A dewatering and recharge system was implemented to lower the groundwater table within the excavation area and minimise the water table drawdown outside the tunnel. Waterproofing and cathodic protection of the tunnel structure were undertaken to meet the long-term serviceability and durability requirements. Screw piles were used under the approach ramps to resist buoyancy from the groundwater. Strict environmental management was employed to manage waste material and contaminants and maintain the ecological environment. Instrumentation and monitoring during tunnel construction had demonstrated satisfactory performance of the tunnel.
Date

ENVIRONMENTAL MANAGEMENT
The Tugun Bypass was regulated by a number of complex Commonwealth, state and local government environmental laws. The environmental impact assessment required prior to approval involved extensive environmental surveys and subsequent route option analysis. Surveys included assessment of meteorological, hydrological, geological and biological aspects of the environment. Ecologically sustainable development was achieved by avoiding, minimising, mitigating, and/or compensating for environmental impacts, in that order of precedence. Three former landfill sites had to be remediated before construction commenced. As the sites straddled the boundary of the road corridor, remedial works were undertaken in conjunction with the Gold Coast Airport. On-site containment of the waste was deemed impractical and waste removal involved excavation and transportation of 11 190 tonnes of material to a

26/11/2006 10

16/12/2006

5/01/2007

25/01/2007 14/02/2007

6/03/2007

26/03/2007 15/04/2007 4

0 Measured at LHS Measured at Centre Masured at RHS RL of Excavation -20 RL of Top of Fill

-10

-2

-30 Predicted 35 mm (Centre) -40 Predicted 43 mm (RHS) -50 Predicted 45 mm (LHS)

-4

-6

-8

-60

-10

FIG 13 - Settlement monitoring results at Ch5718.

10

Melbourne, VIC, 4 - 7 May 2008

13th Australian Tunnelling Conference

RL of Excavation and Top of Fill (m)

Settlement (mm)