Green Park Tube Station is a London Underground (LU) station situated on the north side of Green Park in central London. The station serves as an interchange between the Piccadilly, Victoria and Jubilee Lines.

The purpose of the project at Green Park is to provide the station with step-free access (SFA) between street level and all operational platforms via a new ticket hall extension and lift system. The station, used by circa 60 million people annually, will thus become accessible for Tube passengers with disabilities and passengers with reduced mobility.

Project overview
The works for the excavation and construction of the lift shaft at Green Park were let as a design build contract by main contractor Tube Lines. The winning tender was presented by the contractor, Joseph Gallagher (JGL), with Capita Symonds as the lead design organisation. The specialist sub-consultant for the shaft and tunnel works was Dr. G. Sauer Company. Dr. Sauer was responsible for the design of the sprayed concrete lining (SCL) primary support and for providing construction support services.

The contract was awarded in February 2009 at which point a six-month design phase commenced. The SFA is due to be completed in time for the start of the Olympics in 2012.

The value of the civils contract was GBP 9M (USD 14.57M), excluding the architectural, mechanical and electrical fit-out works.

The shaft and tunnel works involved the design and construction of a SCL ellipsoid shaft, stub tunnel and stair passage tunnel (see Figure 1, above). The shaft will house two 19-person Machine Room Less lifts which will take Tube passengers from the new subsurface ticket hall extension to the Victoria line and the Piccadilly and Jubilee line interchange passageway. The stub tunnel will provide lobby areas to each lift landing and connection to existing circulation and passageway assets.

The ellipsoid shaft consisted of a smaller upper part, which is belled out for the larger lower part. The upper part extends to approximately 10m below the base of the new ticket hall slab with a major axis of approximately 8.6m and a minor axis of 5.6m for excavation. The upper shaft followed a 2m transition before the larger lower part began. This cross section extended uniformly to the shaft bottom at approximately 27m below the base of the ticket hall slab. The excavated major and minor axis lengths of the lower shaft were approximately 10m and 6.2m respectively.

The new lobby areas were formed by the incremental construction of a stub tunnel from the lower part of the shaft. The excavated height and width of the stub tunnel was approximately 11.2m and 9.6m respectively and the total length approximately 6m.

Three connections were required between the stub tunnel and existing passenger tunnels; passage 5/202, and cross passages 4/210 and 4/211.

The construction of the shaft and stub tunnel was located in close proximity to live LU assets. The station had to remain operational throughout the construction phase necessitating continuous monitoring of the existing structures.

Design
The design of the shaft and tunnel works comprised sprayed concrete primary linings (SCL), cast in place secondary linings and sheet waterproofing membranes.

During the design the impact of construction onto the existing Victoria line escalator barrel was a major issue. Containing three escalators (numbers four, five and six), the barrel connects the Victoria line platforms to the existing ticket hall with escalator number four being closest to the shaft and stub tunnel excavation. Hence, escalator number four was anticipated to suffer most from deformations and was put on hydraulic Jacks to counteract vertical deformations.

The use of SCL was deemed to be the most suitable construction option due to its high flexibility in coping with changes in geometry and its ability to allow connections to existing structures. The primary lining design for shaft and stub tunnel consisted of mesh reinforced SCL with lattice girders. Additional steel bar reinforcement was placed in the linings around the future openings. The excavation and support were advanced in typical rounds of one metre.

In the course of the final design Dr. Sauer developed one three dimensional Finite Element Analysis (FEA) using the software package ABAQUS. The model comprised of pre existing cast iron structures and the new sprayed concrete lined structures (Figure 2, above). The model also included the new ticket hall excavation and a concrete collar surrounding the top section of the escalator barrel at its connection to the ticket hall and upper machine chamber.

The model represented an underground volume of 100m x 100m x 42m and comprised approximately 160,000 three-dimensional continuum elements. The elements were kept small particularly at the location of excavations and in close proximity to cast iron structures with characteristic minimum edge lengths of 0.5m. Precise mesh quality checks were conducted during the modelling process to prevent potential numerical instabilities.

The entire structure was located within London Clay except for the uppermost two to three metres which were within made ground. This was removed during excavation of the new ticket hall.

London Clay was modelled following an elastic plastic constitutive material model. The linear elastic material was characterised by a depth dependent Young’s Modulus, linearly increasing with depth and directly proportional to the linear increase of the undrained shear strength. The Mohr Coulomb failure criterion was used to model soil plasticity in undrained soil conditions. To generate the initial stress state a lateral earth pressure coefficient of one was used for London Clay.

The stiffness of London Clay was modelled without direct dependence on the soil’s micro strain or small strain stiffness. However, strain values and strain changes were observed in the model and the suitability of the used stiffness confirmed. The Young’s Modulus used in the model corresponded well with the tested Young’s Modulus at a strain level of approximately 0.1 percent.

Parametric studies of the Young’s Modulus were undertaken in order to generate expected boundary values of deformation and for the evaluation of model sensitivities. The modulus was varied up to (+/-) 50 per cent with respect to the original design value.

The groundwater was omitted from the stress strain analysis of the soil and no pore water pressures and pore water pressure dissipations were modelled. However, groundwater pressures linearly increasing with depth were applied to the secondary lining for the assessment of lining thickness and required reinforcement.

The pre existing cast iron structures of the Victoria line tunnels and the concrete primary and secondary linings of the proposed structures were modelled using structural shell elements (two dimensional elements with a characteristic shell thickness).

Cast iron elements were characterized by the same cross sectional area (and hence same normal stiffness) as the cast iron segments installed at Green Park Station in the 1960s. This resulted in a uniform shell thickness of 0.05m with regard to the Victoria line platform tunnels. The cast iron was modelled to follow a linear elastic material behaviour (E = 100GPa, ? = 0.26). Sprayed concrete primary (E = 15GPa, ? = 0.20) and cast in place concrete secondary linings (E = 26.5GPa, ? = 0.20) were modelled as elastic plastic materials.

The construction sequence was modelled according to the proposed excavation and support methods. This saw the shaft excavation advancing in one metre steps, starting after the simplified modelling of the construction of the pre existing structures and the excavation of the new ticket hall. The stub tunnel, stair passage and connections were advanced continuously in one metre steps representing the designed sequence. In total the analysis comprised 130 individual analysis steps.

At the beginning of the design phase, the construction of a pipe arch over the stub tunnel was deemed necessary to protect the escalator barrel from excessive deformations. The benefit of the pipe arch was analytically investigated in the model and found to be minor. After discussions between Tube Lines and the design build team the pipe arch was discarded in favour of smooth metal spiles.

The FEA deformations taken for comparison in Section Five were ‘best estimate’ results based on the use of the proposed design soil properties. Results obtained from parametric studies deviated up to 30 percent from these results.

Construction
The project’s overall monitoring scheme comprised:
In shaft / in tunnel monitoring,
Ground movement monitoring through inclinometers and
Monitoring of existing assets such as Green Park Tube Station, Ritz Hotel, Devonshire House and Piccadilly Street.

The overall monitoring scheme of existing LU assets, the ground surface and neighbouring buildings was set up by Tube Lines to continuously measure movements, tilts and vibrations.

The Ritz Hotel was equipped with five tiltmeters on the facade and three vibration detectors (Room 121, Casino and Stairwell), and the Devonshire House facade with three tiltmeters. Piccadilly Street and the Green Park site compound were periodically surveyed giving information about surface settlement.

In the Tube station itself the Victoria line platform tunnels, cross passages, the Victoria line escalator barrel, upper and lower escalator machine chambers and the ticket hall were thoroughly monitored. The scheme comprised of 153 prisms, 50 tiltmeters, 88 electrolevels, 37 tape extensometer points and 32 track sensors. All prisms were automatically read from five robotic total stations on an hourly basis. Tiltmeters and electrolevels were read hourly from a data logger. Convergence measurements were performed weekly in the Victoria line cross passages and running tunnels using tape extensometers. Approximately 20m of all four rails located in the Victoria line platform tunnels were monitored with track sensors providing hourly readings.

Except for the monitoring instruments installed in the Victoria line platform tunnels, all instruments were read from the beginning of August 2009. Hence, baseline readings were taken for approximately half a year prior to commencement of the shaft sinking works. The platform tunnel instruments were installed later with initial readings taken at the end of January 2010. Shaft sinking commenced one week before these readings were made.

The client specified four monitoring trigger values (green, amber, red and black) indicating to the contractor actions which had to be taken if the corresponding deformation values were reached.

Comparison of predicted and monitored deformations
Figure 3 shows the vertical deformation history of the escalator located closest to the shaft (escalator number four) and figure 4 of the escalator located remote from the shaft (escalator number six).

In general the predictions from the FEA underestimated the vertical deformations by up to 2mm. All construction took place within the green alarm level zone since the amber alarm level trigger value of 10.1mm was not reached during construction.

Northbound tunnel deformations reached values of 17mm at the time of primary lining completion showing a sudden deformation increase at the beginning of February 2010 when shaft sinking partially excavated the existing tunnel lining. Between February 2010 and the primary lining construction completion in mid April 2010 deformation increased slowly but steadily. Readings taken four months afterwards in August 2010 showed a further slight increase in deformation to 18mm.

Southbound tunnel deformations reached values of 12mm at the time of primary lining completion. Generally the deformation increase was steadier compared to the northbound tunnel due to face subdivision and sequential excavation of the stub tunnel.

During construction some platform tunnel monitoring points exceeded 10.1mm deformation and consequently entered the amber alarm level zone. The red alarm level trigger value of 20.1mm was not reached.

The monitored northbound tunnel deformations exceeded those predicted by up to 7mm. During the construction process approximately a quarter of the platform tunnel’s circumference was liberated from the confinement of the surrounding soil instead of the shaft lining only touching the northbound tunnel as modelled in the FEA.

Further, ring closure of the shaft next to the platform tunnel’s centreline could not be achieved as quickly as planned due to time consuming cast iron break out and removal of an old passageway.

The monitored southbound tunnel deformations reflected closely those predicted in the analysis.

Figures 5 and 6 (left) show predicted Victoria line vertical track displacements against monitored displacements read at the platform edge with construction advance. Each graph shows two lines for predicted displacements indicating the two track rails. Rails located closer to the excavation showed higher displacements (3.5mm; 4.5mm) than those further away (1mm; 2.5mm).

Deviations of monitored displacements from predicted ones amounted to approximately 1 2mm. The maximum vertical displacements were overestimated in the predictions.

Conclusion
The Green Park Station SFA design demanded extensive three dimensional FEA due to Tube Lines requirement for receiving information on how the new works would influence existing LU assets. The geometry and close proximity of the new structures to several existing structures made a complex three dimensional analysis necessary to achieve increased accuracy compared to two dimensional FEA and volume loss approximations.

In general the analysis provided a very good indication of locations where deformations were to be expected and corresponding magnitudes. Comparisons between predicted and monitored deformations typically showed a deviation of up to 2mm except for the NB platform tunnel indicating a deviation of approximately 7mm.

In summary, the performed FEA proved to be a good tool for anticipating expected deformations in London Clay and for dimensioning thicknesses of the concrete shaft and tunnel linings.


Figure 1, Proposed structures for the Green Park lift shaft and stub tunnel Figure 2, Pre existing and proposed structures represented in the FEA Figure 3, escalator number four – predicted against monitored vertical deformations (monitoring values obtained from electrolevel readings below the escalator) Figure 4, escalator number six – predicted against monitored vertical deformations (monitoring values obtained from electrolevel readings below the escalator) Figure 5, Northbound track – Predicted against monitored vertical rail displacements Figure 6, Southbound track – Predicted against monitored vertical rail displacements