Abstract

The NZ Transport Agency’s Waterview Connection project involves the construction of 5 km of motorway to complete Auckland’s Western Ring Route. Half of this new link is twin 14.5 m diameter tunnels constructed by Tunnel Boring Machine with 16 sequentially excavated cross-passages. Over most of the tunnelled length the excavation is through weak interbedded sandstones and mudstones; however, a 150 m long zone of moderately strong volcaniclastic grit was identified during site investigations.

This zone was characterised by steeply dipping, fractured and faulted, higher permeability rock which might allow a direct groundwater-surface water connection to overlying Oakley Creek, and increased the risk of drawdown in overlying contaminated and compressible soils. This zone was identified as one of the highest risk areas in terms of potential adverse effects resulting from groundwater drawdown and ground settlement.

Monitoring during drained maintenance stops recorded inflows to the tunnel excavation of up to 30 L/s with near instantaneous drawdown at distances of almost 200 m, and, air bubbles in the creek coincident with the TBM location confirming the connection between groundwater and surface water. Visual inspections and instrumented monitoring were used to regularly inform drive parameters and negligible adverse effects were recorded as a result of the drawdown during tunnelling. Pre-excavation fissure grouting was undertaken ahead of cross-passage excavations to limit groundwater inflows and drawdown, and monitoring confirmed the effectiveness of this technique in these materials. These proactive mitigation measures meant that negligible adverse effects were recorded as a result of the drawdown.

1 PROJECT INTRODUCTION

The Waterview Connection project involves the construction of 5 km of 6-lane motorway through and beneath Auckland’s western suburbs, linking two existing State Highways to complete a motorway ring route around the city. Half of this new link will be in tunnels and the remaining half comprises surface highways and approach trenches (Figure 1).

The segmentally lined mainline tunnels were constructed using a 14.5 m diameter Earth Pressure Balance Tunnel Boring Machine (“the TBM”). The mainline tunnels are linked by 16 sequentially excavated, nominally 6 m diameter cross-passages.

2 GEOLOGICAL CONDITIONS AND IMPLICATIONS FOR TUNNELING AND EXCAVATION

2.1 Geological Setting

The project is located in the western-central suburbs of Auckland, with the tunnel alignment broadly following the incised valley of Oakley Creek, between the Mt Albert (Owairaka) Volcano to the east and the Great North Road/Blockhouse Bay Ridge to the west (Figure 1).

Figure 1: Location map for the project (modified after France et al, 2011)

Tertiary age, weak sandstones and mudstones of the East Coast Bays Formation (ECBF), form the bedrock in the project area and outcrop in the Great North Road / Blockhouse Bay Road Ridge and locally within Oakley Creek. Compressible Tauranga Group alluvium fills the paleo-valley of Oakley Creek and is encountered within the present-day valley of Oakley Creek and beneath the basaltic lava flows to the east of the Creek.

Mt Albert (Owairaka) Volcano erupted onto a now buried paleo-ridge of ECBF and lava flows in-filled the paleo-valley to the north, west and south. The current position of Oakley Creek roughly follows the western edge of the lava flows. Uncontrolled back-filling of historic quarries within the basaltic lava flows has resulted in pockets of landfill adjacent to the Creek.

2.2 Characterisation of Geological Conditions for Tunnelling

The Well-Connected Alliance developed a 3D geological model from the geotechnical investigation data (over 500 locations) and used it to generate geological long and cross sections to assess tunnelling conditions. Six tunnelling reaches were identified based on predominant lithology, hydrogeology and geological structure. Over most of the tunnelled length (5 reaches) the tunnelling and cross passage excavation is through typical, weak ECBF.

However, within the ECBF are lenses of a stronger, coarse grained volcaniclastic grit (commonly referred to as Parnell Grit). Reach 3 comprises 500 m of interbedded weak ECBF and lenses of the stronger grit, but with a central ~150 m long zone in which the tunnel excavation was wholly within the grit. The rock in this zone was also characterised by steeply dipping joints and folded bedding.

2.3 Characterisation of the Grit in Reach 3

In total 155 in-situ (slug, packer or pumping) tests were undertaken in the East Coast Bays Formation over the wider project area. The geometric mean and median of all tests is approximately 3 x 10^-7 m/s, typical of ECBF, with 85 % of all test results falling in the range of 1 x 10^-8 m/s to 5 x 10^-6 m/s (Figure 2).

Figure 2: Permeability test results from typical ECBF and Reach 3 grit

Whilst a broad trend between permeability and rock strength was identified, review of spatial trends indicated that the highest permeability values were associated with the steeply dipping, openly jointed volcaniclastic grit beds located within Reach 3.

The 41 No. in-situ tests undertaken in Reach 3 indicated permeability values of up to 1 x 10^-4 m/s (Figure 2). Pumping tests indicated peak flows (from a 250 mm diameter well) of up to 16 l/s and sustained flows of almost 10 l/s. Recorded drawdown during the pumping testing indicated a clear SW-NE structural control on the extent of drawdown and the potential for drawdown to extend to distances of up to 500 m.

The locations of tests that indicated high permeability, the observed structural control on drawdown and indications of boundary features in the pumping test all coincide with the mapped extent of thick, channel filling grit (Figure 3).

Figure 3: Inferred extent of high permeability (k) grit within Reach 3 (labelled boreholes are referred to in text or subsequent figures)

2.4 Implications for Tunnelling and Excavation

Defects recorded in boreholes and surface lineaments suggest that Oakley Creek likely follows larger scale geological structures, whilst geological mapping of the creek identified that the steeply dipping grit beds observed at the depth of the tunnel extend to the surface, potentially allowing a direct groundwater-surface water connection to the overlying creek.

Additionally, land-use at the surface of Reach 3 comprises sports fields (Phyllis Reserve) that have been created on a former landfill with no basal liner. Hence any drawdown from tunnelling or excavation could result in contaminant migration, whilst settlement at the surface could allow cracking of the landfill cap and / or ponding of surface water that might result in increased leachate generation.

Reach 3 was therefore identified as one of the highest risk areas in terms of potential adverse effects resulting from groundwater drawdown and ground settlement. The use of an EPB-TBM that could be operated in a closed and pressurised mode, was expected to limit the potential for groundwater inflows and drawdown during tunnelling. However, open drained stops were required for planned maintenance of tail brushes, changing of cutting tools etcetera, and one such stop was proposed to be undertaken immediately prior to entering Reach 3 to allow optimal TBM operation through this zone.

Further, the cross passages were sequentially excavated, and an enlarged (~9 m diameter) cross passage (XP12) was required in Reach 3 to form the tunnel’s low point sump. It was identified during the early stages of the project that pre-excavation fissure grouting would be required at XP12 (and several other cross passages) to ensure safe working conditions and reduce groundwater inflows during excavation.

3 GROUNDWATER EFFECTS DURING TUNNELLING AND EXCAVATION

3.1 Drained Maintenance Stop

Immediately prior to entering the highest risk zone, the TBM was stopped for a planned period of maintenance and replacement of the tail brushes, a critical tool for maintaining face pressure and reducing groundwater inflows during tunnelling. The TBM was stopped on 15^th July 2014 and immediate peak groundwater inflows of 30 l/s occurred at the tunnel face, though these rapidly reduced over a few days to ~10 l/s which was then maintained for the 2 week duration of the stop.

Instantaneous drawdown of 18.5 m was recorded in the nearest piezometer (BH732) located 35 m from the face of the TBM, with drawdown approaching near steady state conditions and a maximum 26.3 m drawdown at the end of the two week stop.

As with the earlier pumping testing, a clear structural control on drawdown was observed, with a near-instant response in groundwater level recorded at distances of more than 200 m away, including at BH529 located on the opposite side of Oakley Creek (i.e. across what might often be considered a “groundwater divide”, Figure 4 and Figure 7).

Figure 4: Extent of drawdown at end of drained maintenance stop (boreholes are labelled on Figure 3)

Recovery of groundwater levels was equally rapid, with most piezometers recovering to 50 % to 85 % of their pre-strop water levels within 7 days, and all fully recovering within 2 months. The only exception being BH529, which some 3 years later has still not fully recovered with an apparent permanent drawdown of 5 m relative to pre-tunnelling conditions.

3.2 TBM Tunnelling Induced Effects on Groundwater and Surface Water

Generally, where tunnelling was undertaken in both the typical ECBF and the more permeable volcaniclastic grit, and with a closed and pressurised face, there was negligible recorded drawdown of the groundwater level.

In some areas, where there was a concern over mechanical ground movements at the surface, the TBM was operated at a higher face pressure in order to allow optimal performance (and hence speed) through the area of concern, while also limiting mechanical movements and groundwater inflows. Reach 3 was identified as one such area as a result of the large groundwater inflows during the maintenance stop, the occurrence of contaminated land fill at the surface, and the close proximity of buildings (with known weather-tightness issues) within 30 m of the tunnel alignment.

As this was also an area of potential connection between groundwater and surface water, twice-daily visual inspections of the creek were undertaken, in addition to continuous stream gauging, in order to visually assess if any irregularities in flow were occurring. During one of these inspections air bubbles were noted rising from the creek bed (Figure 5), coincident with the TBM location, confirming the connection between groundwater and surface water.

Figure 5: Photo of air bubbles observed in Oakley Creek during tunnelling and map showing TBM location at time

The TBM face pressure was slowly lowered until the bubbles reduced to an acceptable level, whilst still maintaining minimal groundwater inflows to the face or drawdown in adjacent piezometers. Daily monitoring of piezometers in the immediate area was then used as a check against the TBM drive parameters (that were set early in the design phase) to confirm the optimal operational pressure.

3.3 Cross-passage XP12 Excavation

Prior to cross passage excavation, narrow (76 mm) diameter probe-holes were undertaken from the completed south-bound tunnel, to confirm geological conditions and allow measurements of groundwater inflows to determine if pre-excavation grouting was required. The recorded inflows to the probe-holes indicated that peak inflows to the larger cross passage excavation of 30 l/s to 80 l/s might occur, with sustainable flows of 10 – 20 l/s (i.e. comparable to that observed during the maintenance stop).

Figure 6: Geological cross section through XP12 (grey line indicates canopy tube, red line indicates probe drilling). NB: ECBF comprises grit (steeply dipping out of page).

During probe-hole drilling, and canopy tube drilling, immediate drawdowns of up to 1.5 m were observed in the ECBF at a distance of 14 m from probe-holes (BH732, Figure 7), with detectable drawdown (0.5 m) recorded at a distance of more than 200 m (DH833 and BH529).

Figure 7: Observed drawdown at select monitoring locations in Reach 3 (distances are relative to outer edges of XP12)

A cementitious grout, targeting larger aperture joints was selected (Maclean et al, 2017) with grouting carried out by a specialist grouting contractor, through grout ports pre-installed in the segmental lining. Having observed the connection between TBM pressure at depth and Oakley Creek, care was taken to limit grouting pressures to reduce the risk of grout being ejected at the surface, with visual inspections of the creek and piezometer monitoring undertaken to confirm.

The cross passages were excavated in a series of short advances with spraying of shotcrete for temporary support. Prior to shotcreting the excavated face was logged by an Engineering Geologist and a visual estimation of seepage was made. Pre-excavation grouting was observed to be highly effective in reducing groundwater inflows. At XP12, groundwater was mainly observed as isolated seepage from occasional joints with the flow estimated to be less than l/s.

The grouting was also effective in limiting the magnitude of drawdown at distance (Figure 6), which, for the 9 m diameter excavation (open and dewatered for 5 months), was not significantly greater, and in some cases was less (BH709b) than observed during short term drilling and testing of the probe-holes. The magnitude of drawdown was also less than observed during the drained maintenance stop and flow gauging of Oakley Creek confirmed no impacts on creek flows.

4 CONCLUSIONS

Site investigations and observations from tunnelling through a discrete zone of moderately strong, jointed volcaniclastic grit, identify that significant groundwater flows and groundwater drawdown can propagate rapidly through the high permeability, but low storativity joints and fractures that often accompany these materials. The extent of drawdown was strongly structurally controlled, with drawdown occurring preferentially in a SW-NE direction, at distances of greater than 200 m and beneath the opposite side of Oakley Creek.

Although no impact on base flows in Oakley Creek was noted, connection to the creek was confirmed by the propagation of air bubbles to the surface. Visual inspections and instrumented monitoring were used to regularly inform TBM drive and grouting parameters to reduce the effects at the surface.

Peak groundwater inflows of 30 l/s and sustained inflows of 10 l/s occurred in open excavations through the grit. However, monitoring confirms that fissure grouting was successful in significantly reducing groundwater inflows and controlling the magnitude of drawdown at distance.

The monitoring results confirm the need for monitoring, and proactive management of groundwater effects, when grit is encountered in deep excavations or tunnelling projects.

5 ACKNOWLEDGEMENTS

The writer would like to thank the NZ Transport Agency and its Well-Connected Alliance partners for permission to publish details of the project, and also Ann Williams for her review of this paper.

REFERENCES

France, S.J., Williams, A. & Cammack, E. (2011) Western Ring Route – Waterview Connection Driven Tunnel: Assessment of Groundwater Effects. Proceedings of the 14th Australasian Tunnelling Conference. The Press, London.

Maclean, H.J., Cartwright, S. and Giauque, A. (2017) Fissure grouting and rock defect characterisation for the Waterview cross passage tunnels. in prep.