1 – Description of the project

The Chelsea to Battersea Project forms part of Cadent’s eight-year regulatory investment plan where over GBP 1bn (USD 1.25bn) is being invested in the North London gas network. Due to the significant regeneration of the Battersea and Nine Elms area, reinforcement of the gas network is required to cater for the growth and demand. The existing gas pipelines crossing the River Thames located above ground on Grosvenor Bridge cannot be utilised for the reinforcement therefore a new pipeline crossing beneath the river is required.

The project involves the construction of two shafts either side of the River Thames, one located within the grounds of the Royal Hospital Chelsea gardens and the other at North east corner of the Battersea Park. A pipeline tunnel crosses beneath the river Thames between the two shafts. Figure 1.1 shows the location plan of the project and the tunnel alignment.

The project organisation included Cadent Gas as the Client, tRiiO (a JV of Skanska and Morrison Utility Services) as the Principal Contractor, Skanska UK Construction as the Contractor, Barhale as the specialist tunnelling and shaft construction sub-contractor. Mott MacDonald provided design services for all disciplines including civil, geotechnical and mechanical/gas engineering.

The 330m-long tunnel has been excavated using an EPBM (Iseki Unclemole TCPID 1800, Figure 1.2) using the pipe jacking technique. The tunnel lining was made of 155mm-thick precast concrete pipes. The external diameter of the tunnel is 2140mm, with axis levels from 77.7mATD in vicinity of the Chelsea shaft to 78.7mATD at Battersea site.

The tunnel acts as a sleeve for the 450mm-diameter gas pipeline, which will operate at up to 7 barg. Once the gas pipeline was installed and tested, the voids in the tunnel and shafts were backfilled with cement-based grout and sand respectively.

The Chelsea shaft was excavated to 30m below ground level (bgl) using a combination of caisson methodology followed by underpinning method. At design stage the caisson method was proposed only for the excavation of the superficial deposits. The excavation method was intended to be switched to underpinning when the London Clay stratum was penetrated, and cut-off was achieved. However, given the significant advantage in terms of the productivity, jacked segmental rings were used as far as ground conditions made it possible, up to 22m bgl. The internal diameter of the Chelsea shaft was 7.5m. The shaft lining was made of 225mm thick precast concrete segmental lining. The shaft was served to launch the TBM.

The same approach was followed at the Battersea shaft. The total depth of the excavation was 32m, of which 18m was constructed using the caisson method with the remaining 14m by the underpinning method. The internal diameter of the Battersea shaft was 6m. The shaft lining comprised a 225mm thick precast concrete segmental lining. The shaft was served as reception shaft to recover the TBM at the end of tunnelling works.

In terms of programme, construction was carried out in two periods due to the time constraints imposed by the RHS Chelsea Flower Show and associated events. The excavation of the Chelsea shaft occurred between January 2017 and March 2017. During the summer months (until the end of August 2017), the site was demobilised and temporarily closed until the activities related to the events concluded. Excavation of the Battersea shaft was carried out from May 2017 to August 2017.

Tunnelling works started in mid- October 2017 and were completed by mid-November 2017.

2 – Excavation/ construction sequence of shafts

2.1 – Caisson Method

The top portion of the shafts were constructed with the caisson method. For this, the precast concrete segmental rings are assembled at the surface and pushed down as the shaft is excavated. There is usually an approximately 50mm over excavation by radius. The void between the excavated ground and rings are filled with lubricants such as bentonite or polymer-based material to minimise friction during jacking operations and to provide stability to the excavated ground behind the rings. After the completion of shaft construction, the lubricant material is replaced with cement grout and the voids behind the concrete rings will be filled to ensure a uniform contact between ground and concrete rings.

Figure 2.1 shows the segmental rings and the hydraulic jacks at the top of the shaft and segmental ring installation. Figure 2.2 shows the steel cutting edge installed at the base of the bottom ring as part of the caisson methodology.

Usually the caisson method is used within water bearing and unstable ground conditions only. Once a stable material with no ground water seepage is reached (e.g. London Clay) the construction methodology is changed to the underpinning method. For the Chelsea to Battersea Pipeline Tunnel project, the Caisson method was used as deep as practically possible. See Figure 1.4.

2.2 – Underpinning method

Once stable ground is reached, the construction method is changed to underpinning method in which the concrete segments are installed as the excavation progresses. Once each ring is installed, contact grouting is carried out to fill the voids behind the concrete rings. Excavation and ring installation resumes after completion of the contact grouting for each ring, see Figure 2.3.

3 – Ground and groundwater conditions

Ground conditions in the area are typical of the London Basin, comprising made ground, overlying alluvium and River Terrace Deposits, overlying London Clay to depth. The thickness of the superficial deposits was observed as approximately 8m at the Chelsea site and approximately 12m at the Battersea site, giving a corresponding top of London Clay at 96mATD and 94mATD respectively. Both shafts were in the London Clay stratum.

Groundwater was encountered within the superficial deposits, but no external dewatering operations were necessary during construction.

4 – Monitoring campaign

The instrumentation and monitoring campaign comprised predominantly of precise levelling points (PLPs) around the shafts, and an automated monitoring system including tiltmeters and hydrostatic liquid levelling cells (HLC) on the river walls. As part of a secondary means of monitoring, prisms were also installed on the river walls and were measured if the automated systems showed any breach of trigger levels. Additionally, nine monitoring manholes were installed to a typical depth of 600mm, near the Chelsea Shaft and along the Chelsea Embankment. These were intended to measure the settlements disregarding the potential stiffness effects of the road pavement.

Surface horizontal movements were not recorded. Figure 4.1 shows the monitoring systems installed.

4.1 – Baseline monitoring

In order to establish baseline levels of movement, the following approaches were taken:

  • For the river walls, automated monitoring commenced three months before start of the construction works.
  • For the ground surface monitoring (PLPs), monitoring started two weeks prior to start of the shaft construction works, with daily readings.

4.2 Monitoring frequency

The frequency of the reading for the monitoring system were as follows:

  • The instruments on the river walls (tiltmeters and HLCs) were automated monitoring system, i.e., real time monitoring system.
  • For the ground surface monitoring (PLPs), the monitoring frequency was minimum daily, which was increased to twice daily if any trigger levels were breached.

After completion of the shaft construction works, all the instrument readings were set back to zero to independently measure the settlements induced during the construction of the tunnel.

5 – Settlements induced by shaft construction

5.1 – Settlements from caisson and underpinning construction methodologies

The development of settlement during shaft excavation has been analysed against time. Three phases of construction have been differentiated:

  • Excavation of the top portion of the shaft by caisson method;
  • Excavation of the bottom portion of the shaft by underpinning method; and
  • Excavation and construction of the base slab. Figure 5.1 and 5.2 show the development of surface settlement during shaft construction. Greater settlements were observed from the underpinning construction method than those observed during the caisson method. In the case of the Chelsea shaft, settlements observed during construction by caisson were around a third of the ones induced by the underpinning method. This was despite the caisson method was used to a depth of 22m below ground surface level (73% of the shaft depth).

A similar trend was observed at the Battersea shaft site, where average settlements from the 18m deep caisson excavation were close to half of those recorded from the underpinning method carried out in the final 14m.

This might be expected as the temporary loss of lateral confinement during the excavation of each ring using underpinning method may allow additional movement and hence volume loss at surface level. This is not applicable to caisson method in which the space behind the rings is protected by a liquid such as bentonite at all time during construction.

After completion of the base slab construction, settlements stabilised and virtually remained invariable.

5.2 – Total settlements – comparison with theoretical predictions

Settlements recorded by the end of the shaft construction and distance from the shaft wall have been normalised by the shaft height (H). Results are shown in Figure 5.3 (pages 33 and 34), which includes the predictions made prior to the construction for comparison. The methodology assumed for prediction is based on New & Bowers (1994).

Maximum settlements in the vicinity of the shaft were observed to be around 0.035% of the height at Chelsea, and around 0.045% at Battersea. No significant settlements were recorded beyond a distance of 1.9×H from the shaft wall. Comparison of movements related to shaft diameter was not possible, as both shafts were of similar size (7.95m and 6.45m OD).

The results have been compared with recent works by New (2017) and Faustin et. al. (2018), where vertical movements have been related to the methodology of construction. According to the research, a pre-installed lining methodology, i.e., sheet pile walls and diaphragm walls, are associated with smaller settlements of around 0.02%, while concurrent shaft linings are thought to induce settlements in the order of 0.06% in agreement with New & Bowers (1994) approach. Construction by caisson has been found to be an intermediate case between the two settlement trends.

The readings from the Chelsea shaft are in agreement with data presented by Faustin et. al. (2018) recorded from shafts entirely excavated by caisson method or by combined caisson and SCL methodology. This seems reasonable as around 73% of the shaft was excavated using caisson method (jacked rings).

The settlements from Battersea shaft however, seem to be larger than the ones obtained from solely caisson shafts, although are in agreement with the dual-lined construction (caisson + concurrent methodology). The proportion of caisson/underpinning excavation depth in this shaft was around 56/44% and it was expected that underpinning would have greater influence on ground movements. This is consistent with the behaviour observed during construction and validates the hypothesis that temporarily unsupported excavation may lead to larger surface settlements around the shaft.

6 – Settlement induced by tunnel construction

The settlements induced by the tunnel construction were monitored from October 2017 to November 2017 when the machine reached the Battersea reception shaft. After completion of the shafts construction and prior to starting of the tunnelling works, the instrumentation and monitoring system was set back to zero in order to differentiate movements caused by the shaft excavation and those from the tunnel excavation.

6.1 – Settlements perpendicular to the tunnel axis

Surface settlements recorded perpendicular to the tunnel were consistent with the expected Gaussian distribution. The volume loss was back-calculated from four sections crossing the alignment. The location of the monitoring sections and the settlement recorded are shown in Figure 6.1 and Figure 6.2.

Maximum settlements recorded correspond to the section closer to the Chelsea launch shaft (monitoring line C1). The maximum vertical movement observed was 4.1mm. It is unclear why the central points of this section showed larger settlements, as these don’t seem to follow the trend of the surrounding monitoring points. Maximum settlements further along the alignment are around 1.4mm at the north area (Chelsea side) and 2.0mm in the south area (Battersea). It can be observed that the range of ground movements obtained were negligible and were too small to have any significant impact on the surrounding third-party assets.

For the back calculation of the volume loss, the Gaussian distribution assumed for analysis corresponded to the approach suggested by O’Reilly and New (1982) with a K value (trough width parameter) of 0.5. A curve corresponding to 2% of volume loss is included in the figures as a reference, as this was the value assumed during the potential impact assessment on the adjacent third-party assets.

The volume loss from monitoring line C1 has been estimated as approximately 2.5%. However, for monitoring line C2, the volume loss estimated from the movements is 1.2% and from monitoring line C3 virtually zero. For the monitoring line C3, in the vicinity of the Chelsea Embankment river wall, it is thought that ground movement has been mitigated by the stiffness of the river wall.

At the Battersea site, the settlements recorded at monitoring line B1 correspond to 2% volume loss.

Volume loss data from pipe-jacked tunnels is limited in the literature and few recent cases are reported with comparable tunnel depths. The values obtained seem to indicate larger volume losses than conventionally expected from TBM tunnelling methods, these could be attributed to some of these factors:

  • The influence of the scale, i.e., the proportion of the tunnel lining OD to the excavation diameter;
  • Contact grouting that is carried out at the end of tunnel construction in pipejacking which allows sufficient time for ground to relax;
  • Jacking forces
  • Face pressure
  • Choice of the machine.

Further research opportunities are open regarding the volume loss as a result of tunnelling by pipe jacking method.

6.2 – Settlements along the tunnel axis

The data that have been recorded along the tunnel axis have been analysed to investigate the development of the settlement trough ahead of the advancing tunnel face and also to track its development after the TBM has passed.

Figure 6.3 and Figure 6.4 show the movements recorded ahead of tunnel excavation. For both the Chelsea and Battersea sites, the settlement trough in front of the tunnel face seems to match the cumulative distribution curve typically assumed during impact assessment practice.

As the machine progresses, settlements are shown to increase as expected to reach maximum values after approximately 40m from the excavation front. Monitoring readings at monitoring lines C2 and C3 (35m to 60m from the Chelsea shaft), are related to backcalculated volume losses from near zero to 1.2% (Figure 6.5).

However, for the points located from 10m to 20m from the Chelsea launch shaft (monitoring line C1), settlements approximately 35m behind the tunnel face correspond to 2% volume loss, but movements continue to develop even after a month, when the tunnel was fully constructed. This behaviour was not observed for the rest of the monitoring sections, and in the transverse direction these monitoring points do not follow the trend of the surrounding points (Figure 6.1, previous page).

For the Battersea site, the recorded settlements also follow the cumulative distribution curve. Back-calculated volume losses of near 1% are estimated by the day of completion of the tunnel, but maximum vertical movements almost duplicate (approximately 2% volume loss) after 10 days from completion of the tunnel excavation (Figure 6.6).

Tunnel excavation works completed on 10/11/2017. Grouting works to fill the tunnel’s annular gap were completed by 17/11/2017.

7 – River walls

Data recorded from instruments installed along the river walls showed negligible movements for both the Chelsea Embankment and Battersea river walls.

At the Battersea river wall, hydrostatic levelling cells showed a settlement trend comparable with the readings from the PLPs. In the case of the tiltmeters, movements of less than 2mm/m were observed during shaft construction and smaller than 0.5mm/m during tunnelling. In the case of the Chelsea river wall, the wall was outside the zone of influence of the shaft construction, and tunnelling has not caused any significant movements in the structure. For both river walls, it is believed that the recorded impacts were mainly influenced by the tidal activities of the Thames and generally followed the seasonal behaviour.

8 – Conclusions

From analysis of the monitoring data recorded during the Chelsea to Battersea Pipeline Tunnel project, the following can be concluded:

  • The methodology of shaft construction plays an important part in the development of settlements around the shafts. Ground movements resulting from caisson construction method have been shown to be smaller than those induced by the underpinning method.
  • When related to the magnitude of the maximum settlements, the combination of caisson and underpinning for excavation of 6.45m and 7.95m external diameter shafts (6.0m and 7.5m ID) has resulted in normalised maximum settlements with shaft depth of 0.045% and 0.035% respectively. This seems to be in agreement with the values reported in the literature.
  • The ground surface settlements recorded from pipe-jacking method were small and follow a typical Gaussian distribution perpendicular to the tunnel axis and a cumulative distribution curve along the tunnel axis. Volume losses have been estimated from 2.5% to virtually zero along the alignment. Further study opportunities are available to investigate the mechanisms leading to the volume loss and how this could be improved in the future.