The project demonstrates the creation of value from a successful basement construction in a difficult urban site. A key challenge on this city centre site was to construct a basement over the LUL Victoria line tunnel within its exclusion zone. This set an engineering design challenge that combined innovative methods of construction, computer modelling of the problem, and negotiation of the solution with key stakeholders. This project innovated in circumstances where heavy penalties for failure were in place to protect the underground network.

Site Setting

The site is in the London Borough of Westminster. It is about 200m to the north of Oxford Circus and bounded by Portland Place to the west and Hallam Street and Great Portland Street to the east. The site’s northern boundary is delineated by Duchess Street, and the southern boundary by All Souls Place. Langham Street bisects the site and forms the boundary between Phases I and II of the development.

The proposed headquarters for the BBC consolidates a number of separate buildings into a unified structure to support their Radio, News and World Service broadcasting services.

The development is constructed in two phases (figure 1) to allow continuous broadcasting operation. Phase I comprises refurbishment and alteration of the existing OBH building which occurred concurrently with the demolition of Egton House and 16-28 Langham Street and redevelopment of this part of the site. The new Phase I building is known as the Egton Wing and comprises a three level deep basement on raft foundation and a five to seven storey above ground structure.

The Egton Wing basement is contained within a secant pile retaining wall on three sides and a sheet pile retaining wall to the north. The sheet pile wall provides a temporary function and has been removed during Phase II of the development to create a unified basement.

Ground conditions and the Victoria Line

The stratigraphy encountered at the site generally consisted of Made Ground overlying Lynch Hill Gravel, London Clay, Lambeth Group and Chalk (BGS, 1994). The Thanet Sand stratum was not identified during the site investigation (FES, 2002). The general geological progression identified (whitbybird, 2002) is summarised schematically in Figure 2.

The results of groundwater monitoring undertaken during the site investigation show that the groundwater level is at +20.0mOD, near the base of the Terrace Deposits. The pore water pressure increases hydrostatically with depth within the London Clay. The pore water pressure identified from piezometers in the Lambeth Group suggest that the stratum is under-drained by the chalk strata and the pore water pressures reduce to 0kPa at the interface between the Lambeth Group and the chalk.

The Victoria line northbound tunnel runs directly beneath the Phase 1 site. The crown of the tunnel is at +9.5mOD (approximately 18m below street level and approximately 5.1m below the bottom of the new basement raft (figure 3). The Victoria line southbound tunnel passes close to the east of 6-28 Langham Street and Egton House with the tunnel crown at a depth of about 24m below ground level.

The northbound tunnel section below the site consists of a knuckle jointed, flexible (i.e. no bolts), cast-iron lining whereas the southbound tunnel is of bolted cast-iron. In both tunnels a bituminous material was applied to radial and circumferential joints in order to prevent water ingress.

Inspections of the north and southbound tunnels were carried out during the summer of 2001. The northbound tunnel generally appeared to be in good condition with limited local signs of minor ingress of water. In the southbound tunnel numerous wet and actively seeping joints were observed throughout the section of expanded lining.

Ground movement as a result of changes in overburden pressure during demolition and excavation represented a significant hazard to the tunnels. The key risks were identified as:

• delays to construction programme due to approvals associated with excavation and construction within the LUL exclusion zones (exclusion zone is defined as 3m to either side of the tunnel and 6m above the tunnel crown)

• tunnel movement causing encroachment into the dynamic envelope of Victoria line tunnel trains in the tunnels and disruption to the operation of the Victoria line tunnel trains

• catastrophic failure by undermining the integrity of the tunnel

• construction activities causing physical damage to the tunnel lining

• train-excited vibrations transmitted into building and affecting broadcasting operation

Managing geotechnical risks

The main activities and deliverables that were undertaken to manage the geotechnical risks during the design and construction of Phase I of the development are summarized in Table 1.

Early on in the scheme design the risks of development on LUL assets were identified. A series of meetings and site inspections were arranged (RIBA Stage B) to discuss the scheme and to identify issues, which could impact the operation and integrity of the tunnels. This was very beneficial and a continuous dialogue was maintained with LUL and their engineers throughout design development.

In order to create the ground model and to be able to assess the impact of development on the tunnels and adjacent structures the site investigation for the works was carried out early in the design phase (RIBA Stage C).

A Conceptual Design Statement (CDS) was written for LUL to ensure that design work for the BBC W1 Project met the LUL requirements for safety and quality and that unnecessary risk was not transferred to LUL (whitbybird, 2002). The report also describes the methodology adopted by whitbybird in their engineering design of the proposed development.

Metronet was commissioned to carry out a line and level survey to confirm the exact location of the tunnels. Gauging surveys were also undertaken to establish the gap available between the sides of tunnel and the train for both Victoria line tunnels. A condition survey was completed and showed that the tunnel was in good condition with some minor water seepages.

Considering the risks to the operation of running tunnels it was recognised that any disruption to the operation of Victoria line tunnels would have had serious consequences on Oxford Circus Station, one of the busiest underground stations in London.

Considering the extent of construction activities in relation to the Victoria line tunnels and in order to minimize the risks and provide an early indication of potential construction hazards on the tunnel it was agreed to monitor the tunnels. The other two main objectives of the monitoring were to: verify design assumption by means of constantly comparing measured and predicted ground movements; and identify and prevent adverse impacts when and if they occurred.

Electrolevels, strain gauges and tiltmeters were installed in the tunnels. The locations of the instrumentation in the tunnel are shown in Figure 4. Survey targets were also installed in the tunnels for back-up manual monitoring.

The electrolevel beams were installed to record a longitudinal heave profile for the tunnel. The tiltmeters record the magnitude of rotation of the individual tunnel segments. The strain gauges similarly record the trend of changes to the magnitude of hoop forces in the tunnel lining. Track surveys were also used as a check on the automated monitoring.

The controls applied to the construction process were determined by Metronet, who prepared an Emergency Preparedness Plan (MRBCV, 2003) that defined trigger levels for movement of the tunnel. The trigger levels are listed in Table 2 and the monitoring equipment was used to verify that deflections were controlled below these limits.

Numerical Analysis

Numerical analyses using the Plaxis software package were carried out to predict the displacement magnitude of the Victoria line tunnels, both during construction and in the long term following project completion. Two sections were used to assess the effects of the redevelopment on these tunnels. The location of these sections is indicated with arrows on Figure 1.

The elasto-plastic Mohr-Coulomb soil model was used to simulate the behaviour of the majority of the strata underlying the site. However, the London Clay was modelled using a higher order (Hardening-Soil) model to simulate the difference in behaviour between primary loading and the subsequent unloading and reloading of the strata. A detailed model of the stress history of the site was prepared to simulate the complex locked-in stresses in the ground.

The results from the Section 1 finite element model were used to predict the longitudinal profiles of the northbound Victoria line tunnel at various stages of construction.

Figure 5 shows the predicted vertical tunnel movement profile at three stages of construction:

• following the maximum excavation depth being achieved (formation level – 14.5mOD)

• construction completed

• long-term conditions (pore-water pressures equalised)

The prediction indicates that, whereas the tunnel inside the footprint of the excavation heaves up to a maximum of 30mm, that outside it settles slightly.

Figure 5 indicates that the analysis predicted about 8mm of tunnel deformation directly beneath the secant piles. Given the lining flexibility in the longitudinal direction, the predicted deflected profile did not give rise for concern. Section 2 modelled both the north and southbound Victoria line tunnels. As with the previous analysis, heave was predicted in the northbound tunnel.

Deformation of the tunnel section was also expected causing stresses in the lining. In general, the lining was predicted to behave in a flexible manner (small bending moments were predicted). The hoop forces in the lining were predicted to drop as the excavation proceeds while the bending moments increase. In the long term, as load is added, the axial forces in the lining increase again, while the bending moments remain approximately the same. Based on the guidance included within LUL Engineering Standard E 3322 A2, Deep tunnels and shafts – assessment, an interaction diagram was produced for the Victoria line northbound tunnel lining (Figure 6).

Moment and axial force predictions for the northbound tunnel are plotted in Figure 6. It can be seen that the plotted values are within the limiting resistance curves of the tunnel lining. The lining structure was therefore not considered to be at risk from the stresses associated with transverse distortion.

Monitoring results

The monitoring equipment was installed in October 2002 and monitoring started on 13 November 2002. This allowed background monitoring to be completed before demolition and piling commenced. A summary of the construction programme is shown in Table 3.

The tunnel monitoring results for the period from pre-construction background monitoring to completion of excavation and raft construction were presented in a final report (whitbybird, 2004). The important conclusions from the report that describe the tunnel movement are discussed below:

Electrolevel Beam 17 at the crown of the tunnel is located at the centre of the excavation and showed the greatest magnitude of heave during the works. The movement trend of this electrolevel beam is shown on Figure 7. This shows minor movement during background monitoring period of 13 November 2002 to 06 January 2003. The demolition of the buildings did not cause any significant movement. The first distinctive movement occurred towards the end of March and early April. This was when sheet piling was installed and water jetting was used to install the piles through the London Clay.

During June 2003, 3m of overburden was removed, reducing the site level to 24.5mOD. The monitoring equipment reflected this unloading with a heave of approximately 1mm. The bulk of the excavation occurred between September and November 2003. The monitoring results show a gradual heave of up to 30mm to 27 November as a further 10m of overburden was excavated over the tunnel. This heave occurred at a rate of approximately 3-4mm per week during this stage of the works. The raft foundation was poured between 27 November and 19 December 2003 and this caused a 1-2mm settlement of the crown of the tunnel. Subsequently, the tunnel continued to heave up to 34mm at a rate of 1mm per week. The remaining raft foundation, basement and ground floor pours over the tunnel started in February 2004 and continued until June 2004. Over this period the crown settled by 1-2mm to 30mm by June 2004. The crown continued to settle in this way during the pours of the 1st floor to 28mm by 23 August 2004. The tunnel shows signs of continuing settlement at the rate of 1mm per month, as more slabs are added. The maximum heave was 34mm. This was recorded during December 2003, when the excavation reached its maximum depth.

Two manual surveys of the tunnel and rails were carried out by Metronet BCV during the works. These visits were made during engineering hours on 8-9 November 2003 and 4-5 December 2003. Both sets of results show that the rails moved by a smaller magnitude than the crown and sidewall of the tunnel. The surveys showed that the rail movement was within the amber trigger level zone for relative track movement between 5m sections.

The peak movement recorded by the live monitoring and manual survey on the dates of the Metronet tunnel visits is shown in Table 4.

The magnitude of movement of the electrolevel beams in the crown of the tunnel remained within the amber trigger level zone for relative track movement between 5m sections.

In general, the profile of movement in the rails and sidewall of the tunnel showed a smaller magnitude of vertical deflection than the crown of the tunnel.

Conclusions

This paper summarised the processes by which geotechnical risks were addressed on a challenging project in a difficult urban environment. In particular the impact of construction activities on the Victoria line northbound tunnel were accurately quantified by numerical analysis and this was later confirmed by continuous live monitoring of the tunnels. The monitoring results have been reviewed consistently since November 2002 and now that construction works have been completed for Phase I, the client no longer sees the business need to continue monitoring. Since the data is so valuable, it was agreed to transfer the package to Southampton University to continue the monitoring, carry out research into the long-term response of the ground and the tunnel and make the data available to the industry. T&T


Plan of original site layout showing LUL tunnels and proposed Fig 1 Geological cross section of the site Fig 2 Victoria line northbound tunnel below Egton Wing substructure Fig 3 Locations of instrumentation in the tunnel Fig 4 Table 1 Table 2 Predicted crown movement of the Victoria Line northbound tunnel during and after construction Fig 5 Moment and thrust diagram for the northbound Victoria line tunnel lining Fig 6 Trend of movement of electrolevel beam 17 at the crown of the northbound Victoria line tunnel. Fig 7 Table 3 Table 4