Blog

Jul 17, 2023

Technical paper: Pile construction using polymer support fluid at Gull Wing Bridge, Lowestoft

By Alex Wallis-Evans and Duncan Nicholson, Arup; Henry Spinks-Essam, KB

By Alex Wallis-Evans and Duncan Nicholson, Arup; Henry Spinks-Essam, KB International; and Steven Heaney, Quinn Piling

The Gull Wing is a new 350m long crossing of Lake Lothing in Lowestoft with a centrepiece 39.5m long rolling bascule bridge, which is opened and closed using hydraulic cylinders (see Figure 1). The bridge deck is up to 22m wide and supports a single carriageway with combined pedestrian and cycle paths on either side. The bridge crossing and approach embankments on either side will link into the existing road network and alleviate the traffic congestion in the town associated with the short opening periods of the existing bridge to allow vessels to pass.

The rolling bascule will be the largest moveable bridge of its type in the world operated by hydraulic cylinders. The client for the project is Suffolk County Council and the main contractor is Farrans Construction. Arup is the lead designer.

The Gull Wing structures comprise a south approach viaduct, a north approach viaduct and a rolling bascule bridge, (Kanaris, 2022). Two of the bridge piers, Pier 4 (P4) and Pier 5 (P5), are situated over water and are supported by 1.08m diameter reinforced concrete rotary bored piles. Quinn Piling was the specialist piling contractor and it in turn appointed KB International to provide its innovative synthetic polymer slurry technology on this project.

The layout of the Pier 4 temporary combi-pile cofferdam is shown in Figure 2. The cofferdam is formed using 1,220mm outside diameter steel tubes which double as temporary casings for the bored pile construction and sheet piles. The combi-wall tubes remain in place to form part of the permanent works bridge piers and support the fender protection.

The ground conditions at the Pier 4 locations have been interpreted from borehole information obtained from the 2019 site investigation. The borehole, BH02A, was completed at the location of Pier 4 and has been used to establish the ground conditions for the rotary bored piling design. Figure 3 shows the interpreted stratigraphy from BH02A.

Figure 3 also shows the standard penetration test (SPT) profile for BH02A. Some of the low SPT results in BH02A may be associated with "blowing sands" encountered during excavation of the borehole. Support fluid was subsequently used to balance the groundwater pressures and stabilise the borehole.

The soil properties adopted for design are summarised in Table 1.

Table 1: Soil profile and properties

Figure 4 presents the grading curves for the Crag Group, showing evenly graded medium sand with typically less than 15% silt and clay and less than 5% gravel. For comparison, the published gradings for tests undertaken on Thanet Sand as part of the Crossrail project in London (Menkiti, et al, 2015) are also plotted. Piles have previously been constructed in Thanet Sand using KB Polymers support fluid and load tests have been carried out as reported by Lam, Troughton, Jeffries and Suckling, 2010.

Lake Lothing is located within the inner Lowestoft Harbour. The water level is influenced by the tide and normal fluctuation is about 2m. It is assumed that the Crag Group is hydraulically connected to sea water level.

The tidal range may be significantly larger at spring tides and during tidal surge events (flooding). In Lowestoft, the high astronomical tide (HAT) level is recorded as +1.48mOD, while mean high water spring (MHWS) is +1.08mOD (National Oceanography Centre (NOC), 2022). Surge tides can also increase the water level. To allow for these effects, the HAT level has been used for assessing fluid level in the pile.

The piles supporting Pier 4 of the crossing are subjected to large static loading when the rolling bascule is in a "closed" position and large cyclical loading as the bridge deck rolls back and forth to permit marine vessels to pass through. The bridge is designed for an average of 10 openings per day throughout its 120 year design life. Therefore the pile design for Pier 4 has considered the axial capacity for both static and cyclic loading conditions.

The designed Pier 4 piles are 1,080mm in diameter, with toe levels ranging from -32.5mOD to -62.5mOD, embedded in the Crag Group. It is not feasible to install full length temporary casing to this depth, therefore the piles were designed as rotary bored reinforced concrete piles to be constructed under polymer support fluid. The design considered the installation of a temporary combi pile cofferdam structure consisting of alternating driven steel tubes and sheet piles to a toe depth of -14mOD prior to the installation of the rotary bored piles. (See Figures 1 and 2). The steel tubes act as a casing to enable the construction of the rotary bored piles. The steel tubes are 1,220mm in external diameter with a wall thickness of 25.4mm. The concrete cut-off level of the rotary bored piles is -5.425mOD.

There was no plan to conduct a construction trial pile or a preliminary/contract pile load test. Therefore, it was crucial that a pile construction method was agreed on with clear monitoring points to check construction quality. Risk assessments were carried out on the pile construction process and mitigations developed. The polymer testing regime was agreed during the production of the piling method statement. The suggested criteria for polymer performance were based on recommendations from Table C20.2 of the ICE Specification for Piling and Embedded Retaining Walls (Sperw), (Institution of Civil Engineers, 2016).

Some of the construction risks are listed below:

The pile designers used pile test results and construction experience from other sites with similar ground conditions to supplement the design. A comprehensive method statement was developed to mitigate the construction risks. Detailed record keeping was agreed during construction to check piling procedures. The allowable construction duration was relaxed from 12 hours after Sperw to 72 hours in the project specification based on previous experience (Lam, Troughton, Jeffries and Suckling, 2010).

To construct the bored piles, a temporary steel deck was installed over each cofferdam on temporary 610mm diameter driven steel piles. The deck level for Pier 4 was +3.0mOD. Top of permanent steel tube level was +2.5mOD. Temporary casings were attached to these steel tubes to extend the top of casing level to +4.2mOD to enable fluid level to extend high enough so that the minimum net fluid pressure could be maintained. This also included an allowance for surge tide effects.

The polymer supplier advised a minimum net fluid pressure of 1.5m above HAT. To allow for the extraction of the drilling bucket a total allowance of 2m was implemented. A compensation tank was used to keep the fluid level topped up during the extraction of the drilling bucket and during fluctuations of the tide. The set-up was arranged so that this could be monitored by observing the fluid level in the compensation tank. Figure 3 shows an illustrative cross section of the polymer set up.

The borehole BH02A identified gravel sized shell layers, therefore KB adopted a high Marsh funnel value to reduce infiltration. The KB Polymer system has additives – Kobbleblok and Magma-fiber – which can be used to reduce flow in permeable soils. KB provided training to the piling operatives and client representatives on the proper use of these additives and also provided guidance on polymer sampling and testing methods during the first week of drilling.

There was a risk that the polymer could become contaminated with sand during drilling. This could lead to a thick filter cake build up on the pile shaft, which reduces the pile shaft friction.

To mitigate this risk, sand content was monitored by sampling and testing the polymer during pile excavation and was kept below a sand content of 0.5%. Likewise, the Marsh funnel values were kept above 90 seconds to minimise polymer flow out of the pile.

The pH was also regularly recorded. Boring was performed very slowly to allow the sand to settle out and minimise stirring sand into suspension. The project specification called for fluid loss tests to be undertaken if sand content exceeded 2% while digging or if the pile exceeded the allowable construction time. The purpose of the fluid loss test was to assess the build-up of filter cake for these conditions.

The pile lengths were up to 60m and this meant that the piles had to stay open overnight and for up to three days. Overnight working was not permitted due to noise regulations. To mitigate this effect, the pile base was kept 2m above the final toe level and cleaned the next day. The drop in fluid level was recorded regularly overnight and topped up as required.

There was very little opportunity to increase the pile length because the Kelly bar was near its full length. Likewise, the pile layout used the cofferdam wall tubes as temporary casings for the piles, leaving only spare tubes for additional piles at the corners of the cofferdam. Additional piles inside the cofferdam would be difficult to install at short notice.

Pier 4 is located about 30m from the shore. This meant that the tremie concrete had to be pumped to the tremie hopper. Tremie concrete was placed using the wet placement method consisting of a vermiculite plug wrapped with plastic tape as recommended in the European Federation of Foundation Contractors and Deep Foundations Institute Tremie Guide (EFFC/DFI, 2018).

The digging bucket had a fluid bypass to avoid pressure reductions below the bucket during extraction. A flat based cleaning bucket was used for the final base cleaning. Two drill holes were burned into the digging bucket to allow the polymer to flow out during the final stage of extraction. Figure 5 shows an example of the digging and cleaning buckets used on the Gull Wing Project. Base hardness testing was undertaken using a 100mm by 100mm diameter steel plate in accordance with Sperw. The plate was also used to record the level of the pile base to check that shaft collapse and the sedimentation of fine material in the drilling fluid was not occurring.

A drilled shaft bottom inspection device (DID) (DMY Inc, 2016) developed by DMY president John Ding was used to inspect the base hardness for the first pile, with an aim to calibrate against the steel plate. The DID could not be used for subsequent piles due to a leakage issue, therefore insufficient data was collected to enable calibration.

Sonic logging was specified on the first five piles to demonstrate the construction method did not lead to significant inclusions in the pile concrete.

Figure 5: Example of digging and cleaning bucket

The results from the routine site records for support fluid are summarised in Figures 6, 7and 8 for Pier 4.

The following points are noted:

During the construction of rotary bored pile P4-7, unexpected weather conditions caused a delay in the digging process. This resulted in the pile bore being left open for 10 days under polymer support fluid, at a depth of -46.5mOD. When piles are left open, a "filter cake" can form due to a build up of sand and silt particles from dirty support fluids on the pile shaft. This is known to be the main cause of pile shaft reduction.

Filter cake build up is more significant with a bentonite drilling fluid compared with a clean polymer fluid and there is little research available showing the impact of dirty polymers on shaft friction. It was therefore very important to show that the polymer fluid remained clean throughout the boring process. A review of available pile load test data generally shows data for piles open for less than a day, with the worst case being left open for 48 hours beneath polymer support fluid, (Lam, Troughton, Jeffries and Suckling, 2010), (Lam, Jefferies and Martin, 2014), (Lam, Jeffries, Suckling and Troughton, 2015).Due to the lack of available data, the pile performance was monitored very closely to ensure the pile bore was kept stable and design assumptions could be made on the impact on the shaft friction.

A second pile, P4-15, was also delayed for an extended period of time. The pile was on standby for nine days at a depth of -45.0mOD and the same observations and site testing methodology was undertaken as P4-7.

Approximately 2kg of Kobblebloc and Magma-fibre were added to P4-7 and P4-15 at the start of the delays as a precautionary measure to seal the Craig Group and prevent any unexpected fluid loss.

The fluid level in the compensation tank was frequently measured so the rate of fluid loss in the piles could be calculated. Figure 9 shows the P4-7 accumulative loss of fluid from the pile to be 12.80 litres/m2 of pile surface over the 10 day period. The P4-15 fluid loss is recorded as 11.30 litres/m2 over the nine day period.

The fluid loss was also used to assess the infiltration distance of the polymer in P4-7. Based on the porosity of the Crag Group, infiltration distance was calculated to be 45mm, and therefore the net fluid pressure hydraulic gradient is steep and sluffing of the pile shaft is unlikely.

Frequent samples were taken and Marsh funnel tests showed that for both piles the polymer viscosity was mostly kept above the lower limit specification requirements of 90 seconds. The viscosity was kept high (>100 seconds) to control infiltration rate and ensure the stability of the bore. For P4-7, total solids content was kept below 1% and is observed to decrease over time due to settling of suspended material. For P4-15, the sand content was slightly higher, reaching a maximum of <1.25%; however this remained within specification limits.

Pile base levels and base hardness readings were taken multiple times a day using the base hardness steel plate. The base level of P4-7 did not change, meaning that sluffing from the sides had not occurred. The build up of debris level at the pile base was less than 120mm for P4-15 over the delayed period, at a rate of approximately 33mm per day, potentially due to the marginally higher sand content falling out of suspension. There were no rapid increases which would reflect collapse of the shaft. This showed that the shaft was stable.

A two hour fluid loss test was undertaken at 100psi for both piles. The measured fluid loss from the P4-7 test was 12.5 litres/m2, which was comparable to the measured fluid loss from the drop in fluid level in the compensation tank, 12.80 litres/m2. The measured fluid loss from the P4-15 test was 9.98 litres/m2, which was comparable to the measured fluid loss from the drop in fluid level in the compensation tank, 11.30 litres/m2. The photographs of the filter paper on completion of the extended tests are shown on Figure 10. These show that negligible debris build up was seen on both filter papers. This gave reassurance that a thick filter cake was not building up on the pile wall.

On P4-7, the results of the site testing and observations gave confidence that there was only a thin build up of filter cake thickness on the shaft and that potential shaft reduction was likely very small. However, as the shaft friction could not be directly measured, and there was no published data to confirm the effect of the long opening time on shaft friction, a conservative approach had to be adopted. It was assumed that 50% degradation of the pile shaft had occurred between bottom of casing -14.0mOD and toe depth during standing -46.5mOD. As a result, the design pile toe was extended by 10m, from -52.5mOD to -62.5mOD. The extended part of the pile was unreinforced.

On P4-15, as per the methodology applied to P4-7, the additional length of pile required for P4-15 was 8.0m. This resulted in a toe level of -70.45mOD. The plant and equipment on site could only reach a toe level of -65mOD without modifications, and therefore it was decided to extend P4-15 and the two adjacent piles to a level of -65.0mOD.

It is evident that polymer technology is evolving rapidly and the use of polymer for piling projects is becoming more common. However, there are still mixed opinions about its use in the industry.

The experience at Gull Wing has demonstrated a method for maintaining a stable pile bore to significant depth, in granular conditions, for a prolonged period of time (>10 days).

Results of the polymer testing show that the sand content in the pile can easily be kept very low and that sand particles settle out of the suspension very quickly, eliminating the need to leave piles overnight to clean the polymer. The results also show that the polymer viscosity degrades with pumping. This could be improved by using positive displacement pumps designed for non-Newtonian shear thinning fluids such as diaphragm, rotary lobe or hose pumps.

At present, while these pumps are available, there is currently a shortage of available and tested solutions on the UK rental market. Diaphragm pumps are available, although they are typically very noisy and do not provide adequate flow rate.

Observations from the polymer sampling indicate that the criteria advised by Sperw can be improved to ensure better performance of polymer fluids. Upper limits on sand content could be reduced and also the upper limit on Marsh funnel time could be increased. The total solids test should be introduced for sand content less than 0.5%.

The fluid loss test is not required for polymers in Sperw and the EFFC/DFI Tremie Guide. However, the test can be used to assess the very thin filter cake thickness as used at Gull Wing.

The pile fluid loss measurements can be used for the calculation of polymer infiltration distance. The very low debris build up on the pile base can be used to show the pile shaft remains stable.

The extended American Petroleum Institute (API) fluid loss test can be used to show that a filter cakedoes not build up over long periods. A lack of filtercake development would result in little reduction of pile shaft friction; however, there is no pile test data to prove this theory.

Review of literature for pile opening times for polymer piles shows a clear gap in research. The industry could benefit from pile load test data for piles in different materials being left open for an extended period. This should be linked to laboratory shear box tests on filter cake material.

British Standards Institution. (2014). BS EN 1997-1:2004 +A1:2013. Eurocode 7: Geotechnical Design - Part 1: General Rules.

Brown D, Muchard M and Bud K. (2003). The Effect of Drilling Fluid On Axial Capacity, Cape Fear River, NC. Deep Foundations, (pp. 30-32). San Diego.

DMY Inc (2016). DID - Ding Inspection Device for Drilled Shaft Inspection. Retrieved from http://www.dmy-inc.com/did.php

EFFC/DFI. (2018). Guide to Tremie Concrete for Deep Foundations.

EFFC/DFI. (2019). Guide to Support Fluids for Deep Foundations.

Institution of Civil Engineers. (2016). ICE Specification for Piling and Embedded Retaining Walls. London: ICE Publishing.

Jardine RJ and Standing JR (2012). Field axial cyclic loading experiments on piles driven in sand. Volume 52, Issue 4, Soils and Foundations, 723-736.

Kanaris S. (2022, November). Gull Wing: Stand Out Structure. New Civil Engineer, pp. 28-29.

Lam C. (2013). Skin Friction Research. University of Oxford.

Lam C, Jefferies S and Martin C. (2014). Effects of polymer and bentonite support fluids on concrete–sand interface. Géotechnique 64, 28-39.

Lam C, Jeffries SA, Suckling TP and Troughton VM (2015). Effects of polymer and bentonite support fluids on the performance. Soils and Foundations, 1487-1500.

Lam C, Troughton V, Jeffries S and Suckling T (2010, October). Effect of support fluids on pile performance -a field trial in east London. Ground Engineering.

Menkiti C, Davis J, Semertzidou K, Abbireddy D, Hight W, Williams J and Black M (2015). The Geology and Geotechnical Properties of the Thanet Sand Formation – An Update from the Crossrail Project. ICE Publishing.

Suffolk County Council. (2023, February 01). Gull Wing Lowestoft. Retrieved from https://gullwingbridge.co.uk/

The National Oceanography Centre (NOC) (2022). National Tidal and Sea Level Facility. Retrieved from Lowestoft Tide Gauge Site: https://ntslf.org/tgi/portinfo?port=Lowestoft

Want to read more? Subscribe to GE's enewsletters and follow us on Twitter and LinkedIn