SCSJV tunnel route connection checks

Matthew Baddeley MCInstCES MRICS, Senior Engineering Surveyor, SCSJV 

Skanska Costain STRABAG joint venture (SCSJV) delivering HS2

SKANSKA COSTAIN STRABAG JV (SCSJV) is delivering full detailed design and construction of the civil works in the southern section of the Phase One project of High Speed 2, UK’s new high-speed railway. The route length being delivered by SCSJV covers just over 26km, with two 21km long and up to 50m deep tunnels running from the London terminus at Euston Station to West Ruislip. In total, six tunnel boring machines (TBMs) measuring 9.8m in diameter will be used to excavate the 21km route.

A seventh, 6.2m diameter TBM was used as part of SCSJV’s contract, to mine a 900m long logistics tunnel that saw a successful breakthrough into Old Oak Common station box in January of this year. This shorter tunnel will eventually allow segment delivery for the Euston tunnels and the removal of muck by a conveyor system.

Figure 1: Leica TS60 total station set up on 88 Wood Lane, sighting towards Westmark Tower, Paddington, London.

Works involving TBMs are always challenging, with a key requirement to ensure the TBM starts tunnelling and finishes in the correct place. A TBM being a long way off the design alignment at breakthrough could cost a contractor millions of pounds to rectify. That is why SCSJV had all TBM excavation routes checked in both horizontal and vertical planes. These 3D connection checks between the primary and secondary control stations were undertaken above ground before any earth had even been broken.

SCSJV understood that if this one-off 8km connection worked, then the below ground tunnel network would be more robust, because of the rigorous control routine that is adopted in the tunnels.Tunnelling is effectively an open traverse that carries a large element of risk when it comes to guiding a TBM along its true 3D tunnel alignment. There are a series of checks that will provide the integrated project team assurance that the TBMs will be set-out in relation to the HS2 survey grid coordinate system^{1 to within a few millimetres. It’s important these activities are scheduled into the main engineering surveying programme:}

1.  Long static GNSS observations for primary and secondary order control networks with a minimum of two independent sessions
2.  Above ground traversing and double levelling connection checks between TBM launch and reception sites
3.  Continual overlap of traverse tunnel measurements, supplemented with optical double levelling
4.  Gyrotheodolite checks
5.  Independent traversing using tripods along the tunnel inverts
6.  Cross passage tie offs between both proposed running tunnels
7.  Potential connection tie offs at shafts prior to breakthrough.

Figure 2: Old Oak Common to London Euston rooftop survey control network.

With consideration of point 2 above, it is common practice to use a traditional optical traverse to check the horizontal accuracy of a tunnel route, with double run levelling to check elevations between portals and shafts. A rooftop traverse was an idea put forward by SCSJV’s head of survey, William Archibald, who had successfully used that method on previous projects.

Three campaigns were used to complete the entire traverse, visiting three buildings at a time; between each campaign techniques were reviewed, revised and re-executed.As the London skyline from Old Oak Common to London Euston has plenty of high-rise buildings, it lends itself to this type of traverse, so the SCSJV engineering surveying team started to plan for a rooftop traverse as far back as November 2020. With help from the SCSJV community engagement team to gain access to properties, the rooftop traverse took a year to plan and execute.

The rooftop traverse is effectively a digital rehearsal that maps the HS2 route above ground. SCSJV understood that if this one-off 8km connection worked, then the below ground tunnel network would be more robust, because of the rigorous control routine that is adopted in the tunnels, together with forced centred brackets in a formation of braced quadrilaterals.

Figure 3: Sighting towards the Trellick Tower, North Kensington, London.

As shown in Figure 2, nine high-rise buildings were selected to traverse an 8km stretch from Victoria Road crossover box in Old Oak Common to London Euston. Horizontal distances measured ranged from 500m for ground to roof level connections, all the way up to 5km.

Three campaigns were used to complete the entire traverse, visiting three buildings at a time; between each campaign techniques were reviewed, revised and re-executed.

A total of 15 SCSJV engineering surveyors were used for each campaign and although it was resource intensive, it was necessary due to the critical nature of this task within the tunnelling programme. It was also important from a health and safety perspective to have a large team because of lone working and to minimise equipment theft.

Due to the redundancy in the rooftop network, the most optimum traverse routes from Old Oak Common to Euston were analysed in LSS using a traditional Bowditch computation.

Three Leica TS60 total stations were utilised for each day’s measure using the following distance measurement modes as a rule of thumb:

• Up to 1.5km: Precise distance measurement mode.
• 1.5 km – 4 km: Once (standard) distance measurement mode.
• Over 4 km: Greater than 4km distance measurement mode.

Ten rounds of face left and face right observations were taken due to the use of manual sighting over 1500m. Because 1º of an incorrect temperature reading equates to 1 parts per million (ppm) error in the field, the onboard total station atmospheric corrections were meticulously checked throughout the campaign. Precise Leica GPH1P prisms were used up to 3.5km and for anything greater than this distance, the Leica GPH3 triple array prism holders as shown in Figures 4 and 7 were used to help catch the EDM signal.

Critically, due to the propagation of angular error over very long distances, it was important that our reference object checks at the end of observations had not drifted by more than 2", otherwise the sets were repeated.

Figure 4: Leica GPH3 triple array prism holder located on pop of 88 Wood Lane, London

 

Due to the redundancy in the rooftop network, the most optimum traverse routes from Old Oak Common to Euston were analysed in LSS (Figure 5) using a traditional Bowditch computation.

The most accurate route was selected and a least squares adjustment was performed in Star*Net using the traverse input TB, T and TE (Figure 6).

Both the traditional Bowditch and least squares misclosures and traverse accuracies were tabulated.

Misclosures at Euston were within acceptable limits for an 8km traverse (Table 1) and most importantly the traverse accuracies were over 1:100,000 (10ppm), which was a key requirement for our above ground tunnelling connection checks.

Not only did the rooftop traverse provide a verification of the relativity of the survey control between Old Oak Common and Euston, but it also afforded a check on the performance of the SnakeGrid projection that has an average fitting of -4 ppm in this area.

Old Oak Common to Euston predominantly runs in an east west direction so one would expect to see an increase in the easting misclosure. This was apparent from SCSJV’s traversing, with magnitudes of 53mm and 35mm attained.

These easting differences are predominately based in a longitudinal chainage direction, so it is considered manageable from a constructability perspective and does not warrant the use of a specific scale factor correction.

Additionally, the SnakeGrid projection was designed to be utilised at track level with a 6.4m elevation difference equating to 1ppm.

Therefore, a small proportion of the misclosure will be attributed to traversing at a significantly higher elevation compared to the elevation that was used in the original SnakeGrid definition.

A major advantage of the rooftop traverse is that it reduced the interface with the general public and vehicles giving benefits not only with health and safety but also limiting the exposure and risk of instrument theft that has become a major problem in the UK engineering surveying industry.

The rooftop method minimised the number of setups compared to a street level traverse that would have incurred numerous total station setups, an increase in error and reduction in accuracy.

The results proved this, with good connectivity seen between Old Oak Common and London Euston.

This has given everybody on the SCSJV integrated project team confidence that the TBMs will excavate in the correct place and tunnelling breakthroughs will accurately hit the bullseye, a key prerequisite for a 360km/h railway.

Table 1: Traverse accuracies from rooftop traversing. 

Figure 5: Initial mean angle reduction and Bowditch traverse computation in LSS.

 

 

Figure 6: Traverse computation in Star*Net and the resulting network plot.

 

Figure 7: Leica GPH3 triple array prism holder located on top of 88 Wood Lane, being observed from The Collective, Old Oak Common, London.

Matthew Baddeley MCInstCES MRICS, Senior Engineering Surveyor, SCSJV

matthew.baddeley@scsrailways.co.uk

scsjv.co.uk

---

1 Turner J, Preston C, Winthrop R, Thatcher I (2020) High Speed 2: Developments in rail engineering survey grids, Civil Engineering Surveyor, online: http://ces.pagelizard.co.uk/webviewer/#cesseptember2020/high_ speed_2_developments_in_rail_engineering_survey_grids