Report on the BGS/ICE Ground Board informal discussion 'Piled Foundations' held at the Institution of Civil Engineers on 21 April, by Claire Watson and Antoine Andrei, Kvaerner Cementation Foundations.
Aspects of piling in challenging ground conditions were illustrated by two presentations which gave the viewpoints of the piling contractor and the geotechnical consultant. Kvaerner Cementation Foundations' contracts manager Tim Fitch opened the lecture by describing the piled foundation and diaphragm walling works carried out on the Eastbourne Marine Treatment Works. Mott MacDonald senior geotechnical engineer William Lau then presented the design and construction of effective piled foundations for the Funchal Airport extension on Madeira. The meeting was chaired by the then chairman of the British Geotechnical Society Bill Craig of the University of Manchester.
Eastbourne marine treatment works
Fitch gave a brief background and details to the project, followed by method of construction and monitoring and testing of the works. The overall success of the project was due to the development of good relations between Kvaerner Cementation Foundations, the client and principal contractor.
Client for the project Operation Sea Clean was Southern Water and principal contractor was Biwater. The site is situated at Langley Point, east of Eastbourne. The original design comprised a subterranean box, 130m by 40m by 15m deep, for a sewage treatment plant to treat 216,000m3 a day. This was to be built, through difficult ground conditions, with 58 T-Panels at a cost of £7.5M.
Kvaerner Cementation Foundations submitted an alternative proposal. This was a circular diaphragm wall with cost savings of over £4M. However this scheme was too radical a change from the existing scheme and consequently was not accepted. At the eleventh hour a hybrid version of the original plan was adopted.
The awarded scheme was a stepped box 15m deep, but shallower at one end, consisting of 25 T-panels, 39 flat panels and four corner panels (Figure 1). There were also 199, 900mm diameter tension piles and 18 1,500mm diameter and 19, 1,800mm diameter bearing piles with plunged columns to an average depth of 50m. The cost of the awarded scheme was £5.5M, with total project cost of £23.5M.
It was difficult to analyse the behaviour of the box using finite element analysis (FEA) due to its stepped structure. FEA performed by Geotechnical Consulting Group under contract to Kvaerner Cementation Foundations showed it was likely that there was a complex interaction between the roof of the structure and the piles and a possibility of heave being generated in the pile which would lead to cracking.
Ground conditions at site were beach deposits, marine sands, complex and variable alluvium and Gault Clay at depth (Figure 2). Stability problems of the diaphragm wall panels were encountered due to the loss of bentonite through the voids in the granular material, so sand was added to the bentonite which filled in the voids in the granular material, preventing further loss.
The sequence of the works was to run piling (Figure 3) and diaphragm walling (Figure 4), concurrently (Figure 5). The box was to be constructed from the deep end at the location of the T-panels. The 28m deep T-panels had a volume of 210m3 and the cages weighed 35t. These were lifted using the tandem technique and carried for 500m to the work site. When exposed, the diaphragm wall panels were within tolerances and were watertight due to the use of a waterbar at panel joints.
When piling at depth, the steel cage was placed under bentonite and the pile was concreted using the tremie technique. It is recommended by the Federation of Piling Specialists to bring concrete to 2.5m above cut- off level but this resulted in large costs in terms of concrete usage and breaking out. A Cemloc device was used to install the plunge columns (Figure 6). By adjusting rollers on the device, greater accuracy for verticality and plan position can be achieved.
There was concern over low overbreak - 3% compared to the expected 7%, so it was decided to expose some piles to find out where the problem lay. It was found that some of the piles were too low mainly because concreting was being carried out under bentonite and it was difficult to determine the level. It then became practice to err on the side of caution and put in more concrete than was necessary.
Two dummy piles were installed remote from the box and tested. A permanent liner backfilled with bentonite was installed and socketed into the Gault Clay. Two tubes were then inserted to allow for extensometers to be placed in the pile with two transducers at cut-off. The load was applied and monitored and the results confirmed design assumptions.
Figure 7 shows the results of the monitoring of the wall one year after completion of the project. Monitoring showed deflection of between 9m and 11m. The point of maximum bending moment as determined by the FEA was also at this depth. Inspection of the wall showed that there were flexural cracks at this location.
Fitch reviewed the technical challenges encountered on the project, quoting mainly the development of the construction process which overcame the difficult ground conditions and the requirement for exceptional accuracy to structural tolerances (Figures 8 and 9).
Funchal Airport extension
Funchal Airport extension is a £250M project to improve the airport facilities on the Atlantic island of Madeira (Ground Engineering June 1998). The existing runway is 1.8km long, only about half the length of the Heathrow Airport runways, and insufficient to receive large Boeing 747 type aircraft. The 900m runway extension will run parallel to the coastline and be realigned towards land to avoid construction over deep water. The new structure will comprise a prestressed concrete deck supported on 3m diameter up to 60m high columns at 32m spacing (Figure 10). To allow work to be carried out on land, a reclaimed platform has been created along the coast using fill obtained from nearby cut slopes. Cost of the foundation scheme is approximately £25M.
William Lau gave a brief description of the project organisation. Within the management team, Mott McDonald was appointed design reviewer for the proposed structures and foundations. The major involvement of its geotechnics division was to check and approve the foundation design. Acknowledgement was given to Beveridge and Rutty1 for their contributions to the project.
Topography and geological setting
The island is located at the top of an extinct volcano, although little seismic activity has been recorded in recent time. The topography is generally rugged with steep slopes, deep valleys and mountain peaks rising to about 2,000m high. Along the coast, volcanic rock has been eroded to form high sea cliffs.
Lau explained the differences between the various volcanic rocks present, ranging from pyroclastic ashes to complex basaltic lava sequences. The two most common rock types on site are lava and pyroclastic deposits. The lava is stronger and have been used as bearing strata for piles. Within the lava, strong basalt up to several metres thick provides the best material for founding piles. However, alternating between the basalt are layers of scoria - a voided material significantly weaker and more compressible than the basalt. Recovery of scoria samples during ground investigation was poor and so the strength of this material could not be tested properly.
A ground investigation was done before construction using rotary coring and the cores were logged by an experienced geologist. The weaker rocks were tested by standard penetration tests, triaxial tests and plate load tests. For the stronger rocks, the core descriptions, rock quality designations (RQDs) and unconfined compressive strengths (UCS) were used to develop rock mass classification to allow assessment of quality and strength.
The ground investigation confirmed the complex and highly variable nature of the geology, as observed in the excavated areas of the site. Locally, the thickness and level of the different rocks can rapidly change over short distances. However, on a wider scale, the lava and the pyroclastic rocks can be seen as volcanic deposits possibly originated from different eruptions forming alternating strata.
In addition, the following underground hazards were considered in the ground model:
Old rivers and beaches with their deposits buried by lava flows. An ancient beach comprising mainly rounded cobbles and boulders was encountered in the lava during piling.
Lava tunnels which would be either empty or filled with loose materials. Two lava tunnels were discovered during general site excavation, one of them was empty and big enough for a person to walk through for over 10m.
Ground conditions are summarised in Figure 11.
Three main types of foundations are being used to support the runway (Figure 12):
Where good quality basalt outcrops, pad footings bear directly on the rock. Settlement is limited to 15mm and, unless footings were located on a slope, bearing capacity is not a problem.
Rockhead drops towards the sea and large diameter bored cast-in-place piles are used to take the loads through the reclamation fill and beach deposits to the underlying rock.
On the slopes above, the selection of a suitable foundation type depended on the combined requirements of slope stability and foundation design.
Pile design parameters
The bored piles are 1.5m diameter and are designed as rock sockets mobilising resistance in the lava. Typical working load is 8MN per pile. Design of the pile lengths considered shaft resistance in the beach deposits and the igneous rocks and base resistance in the rocks. Characteristic shaft resistance was empirically assessed from UCS, with allowable concrete stress as a limit. Base resistance was derived from rock mass classification and a bearing capacity type solution. The lava varies greatly from compact basalt with typical UCS of more than 100MPa to scoria which was too weak to be sampled and tested. Intermediate rock types included weaker forms of the basalt and basalt mixed with scoria.
Improvement of pile design during construction
A total of about 900 bored piles were built, with maximum pile length over 40m. Each column foundation is 12.7m wide and is supported by a group of eight piles. Initially, a borehole was drilled at the centre of each pile group and the design of the piles was based on this borehole data. Because the rock type at the base of each pile was not confirmed, it was always assumed that scoria was present at pile base. This approach was rather conservative since the design relied mainly on shaft resistance and all eight piles had the same length.
During construction it was found that in some cases, even over the short distance between the borehole and the piles, there was poor correlation between the boreholes and the ground observed in the pile excavations around them. Thus, to improve the design for subsequent foundations, boreholes were drilled at the proposed pile positions and were extended below the level of the pile base. This extra information allowed the pile lengths to be adjusted individually so that most piles were end-bearing in good basalt (Figure 13). As a result, the design was able to use more realistic base capacities and the piles became shorter on average. Those extra boreholes were worthwhile considering the cost of piling was about £1,000/m. This demonstrates the benefit of reviewing and improving design during construction.
Construction of piles
Lau presented a series of site photographs to describe the main stages of pile construction:
A casing was installed in the fill and the beach deposits which were drilled through by rotary drilling until further progress became difficult.
Boring was then advanced by chiselling to break up the boulders in the beach deposits and the layers of basalt before removing them by hammer grab.
Reverse circulation drilling also proved to be effective in the more consistent rock such as the thick layers of basalt.
After excavation, the pile toes were cleaned and reinforcement cages installed. The concrete was then placed by tremie and had a high cement content to resist saline groundwater. Each pile had three steel tubes over its entire length to allow integrity testing by sonic logging, with full scale pile tests also carried out to 1.5 times the working load. The results showed that maximum settlements were less than 5mm, well within allowable limits.
Foundations on slopes
The selection of a suitable foundation type to support the columns located on slopes was a major design issue. An example was given for a 12m high, around 60degrees slope, consisting of weak pyroclastic rock. A foundation with typical column load of 65MN was located at a platform behind the slope. The design check was based on Eurocode 7, both design cases B and C were considered to be relevant and had to be satisfied.
Stability of the slope was analysed using the 2-D finite difference code FLAC with the ground being modelled as a linear elastic material with Mohr-Coulomb failure criteria. The first analysis showed that a foundation bearing directly on the platform would not satisfy the design requirements.
To tackle the problem of stability, the feasibility of several options including changing the slope geometry and ground anchors were considered. The contractor preferred using piles as the equipment was readily available and proposed placing mini-piles reinforced with steel sections in a grid (Figure 14). The piles would be 12m long, spaced at 2m centres and rather than being connected to the foundation they would be 'floating' in the ground. The intention was to intercept any potential failure surface that would develop along the slope.
Lau described the interaction between the mini-piles and the ground. Essentially, the ground above the slip surface is moving as a block and the ground below is restraining the mini-piles which are made to act like cantilevers. Based on the work by Viggiani2 on piles used to stabilise slopes, the governing factor for failure is usually the bending moment capacity of the pile section. This was confirmed by the numerical analysis shown in Figure 14; the first row of piles have reached plastic moments just below the slip surface and most of the other piles have developed plastic hinges both above and below the slip surface. The results of the analysis led to the conclusion that the 'floating' piles were not an effective method of dealing with the situation.
The final scheme adopted mini-piles connected to the foundation to take the loads down below any potential slip surface to the underlying lava. Because the steel piles are relatively stiff and the lava relatively hard, the numerical analysis indicated that most of the loads would be carried by the section of pile shaft in the lava, leaving the slope free from foundation load.
Lau said that piled foundations provided a simple, flexible and robust solution to deal with the complex and unfamiliar ground conditions of the remote site.
Alan Powderham of Mott MacDonald commented on the predicted and measured deflections and questioned the water tightness of the joints between the T-panels and the flat panels on the Eastbourne project. He also said that it was unusual to measure cracks in a wall as flexural cracks are difficult to observe. Fitch replied by stating that the water tightness was very good, with just one joint weeping towards the middle of the flat panel section but that the water bar was very effective and there was no evidence that it had come away during the removal of the stop-end. A film of bentonite which would have built up in the joints, only a few millimetres thick, would also improve water tightness. The flexural cracking was an issue as the seepage of water was horizontal and therefore quite unique.
Paul Wheeler, editor and publisher of Ground Engineering, asked Fitch how Cementation Piling constructed its first diaphragm wall in that area in 1969, taking into account the problems that were encountered with the loss of bentonite. Fitch replied that in 1969 the bentonite used was very thick and as Kvaerner Cementation Foundations is now an ISO 9000-accredited company this is no longer acceptable practice.
Bill Craig asked why the circular shaft option, with large cost savings, was not the awarded scheme. Fitch replied that in order to get the same volume, a circular shaft would need to go a lot deeper. This would have meant a radical change to the design of the treatment plant and there was not enough time to rework the scheme. Not only would there need to be changes to the design of the tank but also the component parts, he said. Although it was a very good idea, it was unfortunately too late in the day to change the scheme.
John Sammons, an independent consulting geotechnical engineer, asked if weathering was considered to have contributed to the unusual properties of the scoria at Funchal and, with regard to the piled foundations, how the distribution of shaft resistance for the bored piles would compare with that of the mini-piles used for the foundations on slopes.
Lau replied that the properties of the scoria are mainly attributed to the way the material was formed when the upper and lower surfaces of the lava cooled down quickly. The effect weathering had on its properties is difficult to assess but was expected to be of secondary importance. Regarding shaft resistance for the bored piles, instrumented test piles indicated that shaft resistance developed mainly along the upper part of the pile shaft. This could be due to a combination of reasons including an irregular and enlarged shaft in the beach deposits as a result of the construction method.
1 Rutty P (1998) Bored piles in volcanic rocks for Funchal Airport Extension, Proc 7th Int Conf on Piling and Deep Foundations, June 1998, Deep Foundations Institute, Paper 5.12, p 1-10.
2 Viggiani C (1981) Ultimate lateral load on piles used to stabilise landslides, Proc Xth ICSMFE, Stockholm, Vol. 3, Paper 11/46, p 555-560.