The project
The New Dunedin Hospital (NDH) comprises two main buildings, the 11 storey Inpatient Building and the smaller five storey Outpatient Building, located on two adjacent city blocks on a new site near the existing hospital. With a price tag greater than $ 1 billion, the project is significant on a national scale. Tonkin & Taylor Ltd are the geotechnical engineers for the project, and have been working as part of a team with the structural engineers, Holmes NZ, since the master planning phase in 2018. Te Whatu Ora – Health New Zealand (Health NZ) are the end client for the project.
The site straddles the location of the original shoreline prior to extensive reclamation works in the late nineteenth century. Ground conditions vary as a result, comprising a layer of generally loose reclamation fill several metres thick, overlying soft/loose liquefiable estuarine sediment to a depth of between 5 m and 9 m below current ground level. Dense, alluvial mixtures of gravel, silt and clay underlie the estuarine sediments, and unweathered rock of the Dunedin Volcanic Group was encountered between 40 m and 50 m below current ground level. Due to the proximity of the relatively steep basalt hills of the Dunedin Volcano, the presence of cobbles and boulders within the various layers also adds to the geotechnical challenge at the site.
Figure 1 (1a, top image, site location overlain on Charles Kettle’s 1847 survey map of Dunedin; 1b, bottom image, site location overlain on published geological map, Bishop & Turnbull, 1996)
While a range of raft foundation and ground improvement options were assessed during the concept design phase, the design team arrived at piles as the logical foundation choice due to the high load and performance demands of the buildings (in particular the Inpatient Building, which is to be base isolated), as well as site constraints such as boundaries. Two main pile options were on the table at this concept stage; large reinforced concrete rotary bored piles, and closed-end, bottom driven steel tube (BDST) displacement piles.
Figure 2a (left), bottom driving BDSTs, 2b (right), top driving using a driving frame and bottom driving hammer to facilitate PDA testing of BDSTs
Bored piles were seen as a reliably constructable option, but also inefficient in terms of pile capacity relative to pile size, and thus expensive. Driven piles presented a significant cost saving opportunity having potentially much higher usable pile capacity relative to pile size, but also presented significant construction
risks, specifically:
- Noise and vibration generated by pile driving, effects on the surrounding city environment, and resource consent risk,
- Ability to drive piles into variable ground, including through reclamation fill and cobble/boulder layers,
- Achievable capacity.
To assess the risks and realise the benefits for the project, T+T recommended a full scale trial of driven piles. Health NZ gave the green light soon after the 2020 lockdown.
The pile trial
The pile trial was undertaken in the second half of 2020 and included a total of five 710 mm diameter BDSTs, at three different locations across the NDH site. Large bored piles were not included in the trial, due to the expected lower construction risks relative to the much higher costs to construct and test bored piles. March Construction were engaged to install the trial piles, and proposed the inclusion of ‘Screwsol’ piles as a potential low-noise, low-vibration alternative to driven piles. Screwsol piles are constructed using an auger to displace soil and place concrete via the auger stem. This method is similar to CFA piles, except soil is displaced laterally by the auger rather than removed. Screwsol piles also incorporate a threaded shaft to further enhance shaft friction capacity. Six 530 mm diameter Screwsol piles were installed and tested adjacent to the BDST piles, providing a useful comparison with an alternative low noise and vibration option.
Figure 3a (top), March Construction’s ‘Screwsol’ auger, 3b (bottom), Screwsol pile following insertion of the reinforcing cage
Pile driving analyser (PDA) testing was used to assess the as-built capacity of both pile types, and extensive monitoring was undertaken to measure noise and vibration levels, including attenuation with distance from the works.
A summary of the trial piles and the capacities achieved is presented in the table below.
Parameter | 530 mm diameter Screwsol piles | 710 mm diameter BDST piles |
Pile depth | 10 – 14 m | 11 – 21 m |
Shaft friction capacity | Up to approx. 150 kPa | Up to approx. 200 kPa |
End bearing capacity | 3 – 9 MPa | 17 – 18 MPa |
Geotechnical ultimate pile capacity | 1.4 – 3.5 MN | 8.5 – 9.5 MN |
Project benefits and reflections
The BDST pile capacities exceeded expectations by a significant margin. Without the benefit of the pile trial, the design would have been based on a more conservative assessment of pile capacity using correlations with investigation data as per common practice, so the trial resulted in a much more efficient pile design making use of the available pile capacity. In the case of the Outpatient Building in particular, this resulted in a final design featuring one pile per column, and a large saving in both pile costs as well as at-grade structure costs such as pile caps. While the resultant design is less conservative, this reflects the higher degree of confidence obtained from site-specific trial pile data. Verification of pile capacity using driving data and PDA testing of production piles adds further confidence in the works.
Screwsol piles were proven to be a constructable, low noise and low vibration pile type. In this instance, the lower capacity of Screwsol piles relative to BDSTs meant that a much greater number of piles would be required, as well as associated at-grade structure such as pile caps or a raft. While this option was not selected for the main buildings, Screwsol piles may still be used for associated structure foundations such as oxygen tanks, especially in sensitive areas where noise and vibration limits are critical.
Obtaining resource consent for driving large piles in central Dunedin required a rigorous assessment of the effects of noise and vibration, particularly at the adjacent historic Central Fire Station and Allied Press buildings. The site-specific noise and vibration data obtained from the pile trial showed that the effects could be managed, resulting in achievable and pragmatic resource consent conditions.
If large rotary bored piles were used in place of BDSTs (for example due to noise and vibration constraints or uncertainties around constructability), piles would have to be in the order of 1.2 – 1.5 m diameter and potentially extend to bedrock, resulting in a much larger volume of concrete and steel and therefore a much greater cost and increased carbon footprint. The efficiencies of BDSTs are due to a combination of a) higher shaft and end bearing capacity achieved by driven displacement piles, b) site-specific shaft and end bearing capacities proven during the pile trial, and c) use of more favourable strength reduction factors due to verification of pile capacity using driving data and PDA testing. The resulting foundation design comprises 78 piles supporting the Outpatient Building, and in the order of 300 piles supporting the Inpatient Building.
Figure 4 – Outpatient Building piles under construction
Figure 5 – Outpatient Building structural system, courtesy of Holmes NZ
Clearly a full scale pile trial is not feasible for most projects. The NDH project provided a unique opportunity creating sufficient time for a pilling trial in parallel with demolition of the previous buildings. The scale of the project meant that savings from an efficient pile design would offset the cost of the trial along with increased confidence in a foundation system which could be consented.
As of September 2022, the majority of the 78 piles supporting the Outpatient Building have been driven, and design of the Inpatient Building is ongoing. The pile trial also has the potential to benefit future buildings forming part of the wider health campus and in wider vicinity.