Introduction In the urban environment, pile foundations are frequently constructed in locations very close to existing tunnels. Tunnels carrying transport networks, moving walkways, escalators and services can often tolerate only minimal movements. Construction and loading of piles causes ground movements and stress changes that can adversely affect the tunnels. Tunnel owners, such as London Underground (LUL) therefore place restrictions on the construction of deep foundations close to their tunnels. The restrictions may consist of one or more of the following: maximum allowable tunnel deformations; stress changes in the lining; or clear distances between the tunnel and the piles (Chudleigh et al, 1999).
The problem of pile-tunnel interaction was recognised as early as the 1940s, when pile foundations for the Royal Festival Hall on the Southbank in London were rejected partly because of the possibility of vibrations causing damage to the adjacent tunnels (Measor & New, 1951).
Over 30 years ago, during the construction of the Victoria line on the LUL network, the problem was of major concern and Morgan & Bartlett (1969) stated that 'These [multi-storey buildings] require bored piles up to 6ft [1.8m] in diameter carried down to the level of the underground railways. London Transport is very concerned about this because of the possibility of distortion and damage to its tunnels.' However, the actual impact of piles on tunnels is still not well understood and there are very few published case records (Benton & Phillips, 1991 and Chapman et al, 2001).
This paper presents a case study in which the influence of the construction and loading of bored pile foundations on existing LUL tunnels was measured. This work is part of an Engineering and Physical Sciences Research Council (EPSRC) funded research project in collaboration with LUL and Geotechnical Consulting Group (GCG) in which the interaction problem is investigated by means of field measurements and finite element analyses. The finite element analyses are not presented here.
The Effra riverside housing development has a varying number of storeys, with seven at the river bank and 15 furthest away from the river (Figure 1). The site is immediately to the south of Vauxhall Bridge on the River Thames in London, lying above Vauxhall station on the Victoria line on the LUL network. Higgins et al (1999) describe finite element analyses carried during the project design phase.
Figure 2 shows the foundation layout of the part of the development relevant to this study. The location of the instrumentation, which is discussed in detail below, is also shown.
The LUL tunnel is part of the Victoria line which was built just over 30 years ago. It has an axis depth of -12.35m OD, 17.85m below ground level.
The station tunnel and the running tunnel both have rolled segmental cast iron linings with diameters of 6.9m and 4.076m respectively.
The bored pile foundations are situated in rows running roughly parallel to the LUL tunnel with a centre to centre spacing of about 7.2m.
The two piles most relevant to this paper (P4 and P27) have diameters of 1.8m and depths of close to 60m. The clear distance between the tunnel and piles P4 and P27 is 2.5m.
Figure 3 shows a cross section of the site (A-A through P4 and P27), with ground level at +5.5m OD. The stratigraphy, typical for central London, consists of made ground and Thames Gravel overlying London Clay. The Lambeth Group lies beneath, consisting of alternating layers of clays and sands. The Thanet Sands lie below this.
The water table is at approximately +1m OD, the interface between the made ground and the Thames Gravel. Boreholes in the upper Lambeth Sand (between -40m and -43.5m OD) recorded head levels of -15m and -18m OD. This clearly indicates an under-drained sub-hydrostatic pore water pressure profile, which is typical of central London (Simpson et al, 1989).
Instrumentation LUL often requires developers working close to its tunnels to monitor the tunnel lining to ensure safe operation of the railway during installation and loading of bored pile foundations. For this research, as well as tunnel lining monitoring, new instrumentation was designed to measure soil movements between the tunnel and the piles.
The responses of two tunnel lining rings were monitored using vibrating wire strain gauges. Figure 2 shows the locations of Ring A (in the station tunnel) and Ring B (in the running tunnel).
This paper is concerned with the results obtained from the instruments in Ring A. Figure 4 shows the location of the strain gauges around the tunnel lining, labelled VW A1 to VW A5.
Rod extensometers were used to measure horizontal movements of the soil approximately 0.5m from the pile relative to the tunnel lining (locations are indicated on Figure 2). For this site an instrument was designed that consists of an extensometer rod (inside an outer tube) anchored in the soil, with an electrolevel tip and connected to the tunnel lining with a displacement transducer (Figure 5).
Vertical movements of the soil were measured using strings of four electrolevels installed in a horizontal hole from the tunnel lining towards the piles. The locations of the electrolevel strings are indicated on Figure 2.
An enormous amount of data was generated at the site. This paper focuses on the results from the rod extensometer close to P27 (the only one installed prior to pile construction) and VW A2 and VW A5, the strain gauges closest to P27 and P4, respectively. VW A1 was damaged during the installation of the soil instrumentation and was only replaced after the construction of P27. No results from the electrolevel strings are presented because they were not installed early enough to have given a reliable response to the construction of P27.
Results are presented in two parts.
The response to pile construction
The response to pile loading
The construction sequence for P27 is summarised in Table 1. P27 was cased through the permeable made ground and Thames Gravel. The excavation was then carried out without bentonite support through the London Clay to a depth of -30m OD. Before proceeding into the more permeable Lambeth Group deposits the pile bore was supported by bentonite. The excavation then proceeded to the pile toe level of -54m OD. The bentonite was de-sanded prior to concreting.
Figure 6 shows horizontal movements measured by the rod extensometer plotted against time during the construction of P27.
Positive values indicate movements of the soil at the location of the electrolevel tip, 0.5m from the pile face, towards the tunnel lining. The following key features can be identified and are closely related to the construction activities:
31 January 1999, 12:00 (T1): After T1 a substantial movement occurred away from the tunnel. This can be interpreted as the time at which the excavation of the pile shaft passed the tunnel axis. As dry excavation proceeded this movement continued at a reduced rate up to a maximum of about -5mm.
2 February 1999, 08:00 (T2): A reversal in the movement of 0.75mm was observed. This can be attributed to the introduction of bentonite. Subsequently movements were relatively small with a tendency towards further movements away from the tunnel lining.
During this period the excavation proceeded to the final pile toe level.
l2 February 1999, 19:00 (T3): Movement towards the tunnel lining was measured. This 4mm of movement corresponds to concreting.
2 February 1999, 23:00 (T4): This time corresponds to the end of concreting. The accumulated movements had reduced to less than -1mm.
The movements thereafter were probably related to curing processes of the concrete.
Figure 7 shows the response of two vibrating wire strain gauges in Ring A (VW A2 and VW A5) plotted against time. Arrows indicate the construction periods of P27 and P4.
VW A2 is located on the side of the tunnel close to P27. It can be seen that compressive strains (negative) of approximately 25meare recorded in response to pile excavation. The concreting of the pile causes tensile strains of close to 40me The response of VW A2 after the end of pile construction still shows some significant variations.
VW A5 does not show a response to the construction of P27, but does respond to the construction of P4. The pattern of response is similar to that of VW A2 in response to the construction of P27, ie compressive strains during excavation and tensile strains during concreting. The magnitude of the response, however, is only in the region of 10me The strain gauges are located on the extreme internal fibre of the tunnel lining segments. Therefore, compressive strains can be interpreted as local outward movement of the tunnel lining, whereas tensile strains can be interpreted as local inward movement. This means that during pile construction the tunnel lining is locally moving towards the pile bore, whereas during concreting the lining is moving back.
The measurements show that during bored pile construction the lining only responds on the side on which the pile is constructed.
Figure 8 shows horizontal movements measured next to P27 plotted against time. The horizontal movement on 1 April 1999 was taken as a datum. It can be seen that from the beginning of May until the end of September a reduction of the distance between the pile and the tunnel of nearly 2.5mm was recorded. Subsequently there are hardly any changes until the end of October. The movements correspond to the construction activities. Construction of the superstructure began in April 1999 and was largely completed by the middle of October 1999.
Figure 9 shows the response of vibrating wire gauges VW A2 and VW A5 plotted against time. The readings on 1 April 1999 were again taken as a datum. Both strain gauges show the development of tensile strains over the first six months. At the beginning of October VW A2 measured a tensile strain of around 70mewhereas VW A5 measured 30me In October the measurements of VW A2 level off whereas the ones of VW A5 continue to increase slightly. The time over which these changes were measured also corresponds closely to the construction sequence described above. The magnitudes of these strains are greater than those recorded during pile construction.
Conclusions Measurements of soil and tunnel response to adjacent bored pile construction and loading have been presented. It has been demonstrated that both pile construction and loading can impact on nearby tunnels. In particular:
The rod extensometer picked up horizontal soil movements close to P27 that clearly identified key features of the pile construction process.
The vibrating wire strain gauges on the tunnel lining similarly recorded responses to construction, but only on the side of the tunnel where the pile was located.
The ground movement and lining strains recorded during pile boring were almost returned to zero during concreting, giving a small net effect.
The magnitude of change recorded by the strain gauges during pile loading was much greater than during pile construction. However, the rate of change during pile loading was much smaller than during pile construction.
The measurements taken for this case study give an invaluable insight into the impact of newly constructed bored pile foundations on existing tunnels. As well as providing valuable information themselves, the measurements are being used to validate finite element analyses of the interaction problem. These analyses allow further investigation of the mechanisms of interaction.
The author is grateful to Dr Trevor Addenbrooke and Professor David Potts for the supervision of the research project. Thanks are also due to Ivan Chudleigh from LUL for making the site available for the installation of the instrumentation and Gerwyn Price of CMCS and Jon Austin of LUL who installed the instrumentation and provided the data.
The piling contractor Amec Civil Engineering and the developer St George South London are acknowledged for providing information on the construction details of pile P27 and the superstructure. The author would also like to thank Kelvin Higgins of GCG for his continued interest and support for this project.
Benton LJ and Phillips A (1991). The behaviour of two tunnels beneath a building on piled foundations. Deformation of Soils and Displacements of Structures. X ECSMFE, Florence, Vol 2: 665-668.
Chapman T, Nicholson D and Luby D (2001). Use of the observational method for the construction of piles next to tunnels. Proc Int Conf Response of buildings to excavation induced ground movements, London.
Chudleigh I, Higgins KG, St John HD, Potts DM and Schroeder FC (1999). Pile-tunnel interaction problems. Proc Tunnel Construction & Piling '99, London, 172-185.
Higgins KG, Chudleigh I, St John HD, Potts DM (1999). An example of pile tunnel interaction problems. Proc Int Symp Underground Construction in Soft Ground, IS-Tokyo 99.
Measor EO and New DH (1951). The design and construction of the Royal Festival Hall, South Bank. Journal Instn Civ. Engrs, Vol. 36: 241-318.
Morgan HD and Bartlett JV (1969). The Victoria Line: tunnel design. Proc. Instn Civ. Engrs, Supplement, Paper 72703: 377-395.
Simpson B, Blower T, Craig RN, Wilkinson WB (1989). The engineering implications of rising groundwater levels in the deep aquifer beneath London. CIRIA, Special Publication 69.