The choice of tunnelling as a method of cable installation is becoming more popular in congested urban environments as a means of minimising disruption. The initial cost of tunnelling is higher than traditional trenching techniques, with the construction programmes being of a similar duration. However, substantial benefits can be realised, which include: reducing the risks of clashing with existing services and the time taken for diversions; enabling cable maintenance to be undertaken; reduced whole life costs; cable replacement; allowing the installation of greater lengths of cables with a reduced likelihood of mechanical damage; and the option of installing additional cables.
A recent example is the 2.8m dia. cable tunnel driven from an electricity sub-station in West Ham to the Greenwich Peninsula in east London. The tunnel will contain electricity cables providing power to the Millennium Dome site. Another reason for choosing the tunnel option was the presence of four major obstructions along the route: the River Thames; its tributary, the River Lea; the Docklands Light Railway; and the A13 road.
Before construction started, a monitoring site was established along the route. Its purpose was to investigate the relationship between tunnel construction and the response of London Clay. Several locations were considered for this purpose, but the only feasible site was in a disused timber merchants’ yard 400m from the start of the drive.
Instrumentation was installed in boreholes above and besides the tunnel alignment. Some of the instruments are located within 500mm of the anticipated position of the tunnel lining. At this location the tunnel axis passes at a depth of approximately 13.5m below ground level (-11.5m AOD).
This paper provides a case history of the vertical displacements experienced in the ground as the tunnel passed through the monitoring site, giving an insight into the development of the idealised settlement trough.
Borehole layout and instrumentation
Nine boreholes were sunk in three arrays perpendicular to the tunnel alignment in order to facilitate installation of the sub-surface instrumentation and allow a comprehensive site investigation to be undertaken. The borehole layout is illustrated in Fig 1, which also shows the positions relative to tunnel chainage. Boreholes 2 and 2A are located directly above the tunnel centreline, with the remainder at an offset of 2m in plan either side of the centreline. Each borehole is finished at ground level with a secure manhole installation to prevent vandalism.
Boreholes 1A, 2A and 5A contain multi-point extensometers connected to vibrating wire displacement transducers housed in the reference head, located about 500mm below ground level. Three point extensometers were installed in boreholes 1A and 2A, the lowest anchor point being about 600mm above the tunnel crown and the highest just below the surface of the London Clay. A five point extensometer was installed in Borehole 5A, the lowest anchor point passing down beside the tunnel alignment at an offset of 500mm from the tunnel lining.
Boreholes 1, 2 and 3 all contain earth pressure ‘spade’ cells and piezometers, each instrument incorporating vibrating wire pressure transducers to allow real-time monitoring. The spade cells in boreholes 1 and 2 are located about 600mm above the tunnel crown and the cell in Borehole 3 beside the tunnel lining at axis level. Typically, the piezometers are located about 750mm above the pressure cell.
Inclinometer tubes were grouted into boreholes 4 and 5, passing down either side of the tunnel alignment at an offset of 500mm from the tunnel lining. The tubes terminate about 3m below the tunnel invert. A standpipe piezometer was placed in Borehole 3A to measure the piezometric pressure in the underlying Woolwich & Reading Beds (Lambeth Group).
Before the arrival of the tunnel at the monitoring site, the vibrating wire instrumentation was logged hourly. From January 14 1998, the frequency of logging was increased to every 15 minutes, thereby recording a greater level of detail as the TBM passed through the site in the early hours of January 20 1998. This logging frequency was maintained until January 22 1998, at which time logging reverted to hourly readings.
The datalogger was taken off-hire at the end of March 1998 and, since that date, manual reading of the vibrating wire instruments has been undertaken using a hand-held readout unit at intervals of about three months.
The inclinometer tubes and precise levelling points were all surveyed manually, initially at daily intervals as the tunnel approached the site and four hourly as it passed through. The frequency of these manual readings then reduced to daily and finally to weekly, terminating five weeks after the arrival of the tunnel. One further survey of the inclinometer tubes was carried out five months after construction of the tunnel. Precise levelling continued for a further five months to the end of October 1998.
Geology
The tunnel was driven in London Clay for its entire length. The geology at the location of the monitoring site is shown in Table 1.
In-situ strength and stiffness: standard penetration tests and unconsolidated undrained triaxial tests (UUT) were carried out in conjunction with Atterberg limits and moisture contents to obtain index properties for the London Clay). When combined with shear strengths obtained from the self boring pressuremeter (SBPM) tests, a design line for shear strength in the mass was obtained. At the location of the tunnel axis, this equated to approximately 180kPa.
Tunnel construction and geometry
The tunnel was constructed by J Murphy & Sons in alliance with London Electricity and Brown & Root. It was driven using a Lovat M116SE known as ‘Dorothy’, which is capable of operation in closed mode through difficult ground. The machine had been used previously on projects such as the London Water Ring Main, the North Western Storm Relief Sewer and a number of other tunnelling projects in London Clay.
The TBM had a face diameter of 2.99m with an overcut to 3.004m, giving an average grouted annulus of 102mm between the bolted segmental lining and the excavated bore. The lining comprised 180mm thick reinforced concrete segments with an o.d. of 2.8m, constructed inside a tailskin with an o.d. of 2.9m. The overall length of the TBM and tailskin was 6.9m.
A typical excavation and construction sequence involved the following processes:
This construction sequence was repeated in cycles of approximately half an hour, giving an average rate of advance of about 2m/h, continuing for much of the standard 10h shift. The 2h period between shifts was used for maintenance purposes.
Displacement response of London Clay
The response of the ground to underground construction like tunnelling is generally the formation of a settlement trough above the tunnel. For a single tunnel, this can be idealised, at any elevation, to the shape of a normal distribution curve, located perpendicular to and symmetrically placed about the centreline of the tunnel.
For immediate displacements, the volume bounded by the curve is equal to a notional volume loss, expressed as a percentage of the tunnel face area per metre run of tunnel. Together with a trough width parameter, i, which is proportional to the depth of the tunnel axis below ground level, the curve is mathematically defined and can be used to estimate settlements at any offset from the centreline of the tunnel.
In the case of full face TBM tunnelling, the volume loss can be associated with face loss experienced at the cutting face, and radial closure around both the TBM and the tunnel lining as the ground closes in. It is the progressive development of this setlement trough that is considered further.
Measured vertical displacements near the surface: vertical displacements were measured near the surface by precise levelling of the concrete base slabs in the manhole chambers. Levelling points approximately 0.5m below ground level were used to reduce the influence of the concrete slab that covers the site.
Surface settlements at the location of all the boreholes were first observed at about 9m (three diameters) in advance of the face of the TBM. A graph illustrating surface settlements at boreholes 1, 2 and 3 is shown in Fig 3. The vertical scale shows vertical displacement in millimetres and this is plotted against the distance of the face of the TBM from the monitoring position in metres. A negative value on the horizontal scale, therefore, shows that the face of the TBM is approaching the borehole; a positive value shows that the face of the TBM has passed by.
As the TBM approached the borehole positions, the rate of settlement was approximately uniform, increasing as the TBM reached the instrument position. Once the face of the TBM arrived at the location of the boreholes, approximately 10-15% of the observed maximum immediate settlement had occurred.
The next phase of settlement happened when the face of the TBM was between the instrument position and 7m past the borehole, and accounted for a further 25% of the total observed immediate settlement. Once the trailing edge of the tailskin passed the instrumented borehole, a rapid increase in the rate of settlement was observed. This accounted for the remaining 60-65% of the total immediate settlement and continued for a period corresponding to about 13h after the face of the TBM first arrived at the instrumented borehole.
For this tunnel, using a full face TBM and progressing at a rate of 2m/h, it is evident that settlement in advance of the TBM is a small proportion of the total immediate settlement. Most of the immediate settlement, 25% over the body of the TBM and 60% to 65% behind the tailskin, is directly related to the geometry of the TBM and the construction method used.
The rates of surface settlements are summarised in Table 2 for both vertical movement ahead of the face and once the TBM/tailskin has passed. It should be noted that the rates are much greater than those previously reported3 The most likely reason for this phenomenon is the frequency of measurement. The rates presented in Table 2 are averaged from readings taken at 15 minute intervals over periods of between 80 minutes and 320 minutes. Had the readings been taken at, say, 12h or 24h intervals, the average rates would reduce as a consequence.
The maximum immediate settlement was followed by a short period of heave or swelling, possibly related to grouting the void between the ground and the tunnel lining, and associated with an increase in porewater pressure. Settlements, once again, reached the value of the maximum immediate settlement at approximately two to three days after the passage of the TBM.
Back analysis of the immediate surface settlement trough using the methodology illustrated in Fig 2 indicates a volume loss of 2.3%, with corresponding trough width parameters at the surface of i = 5.2m and k = 0.4. These represent composite values for the clay and the overlying gravel, alluvium and made ground.
Measured sub-surface vertical displacements: When considering greenfield settlement troughs idealised to a normal distribution curve, the volume of ground loss is assumed to be equal at any level between the tunnel crown and the surface. Because the trough width is related to height above the tunnel (i = k.zo), the settlement trough becomes narrower closer to the tunnel crown, (reduction in zo). Consequently, the value of maximum settlement must increase. Further analysis has also shown that the k-value increases with depth similar to the model described by Mair et al².
As expected, the rate of settlement in advance of the TBM was greater close to the tunnel than at the surface. As the TBM passed under the array, settlement virtually ceased at all levels above the tunnel, while the body of the TBM provided ground support. Directly above the crown, 80% of the total immediate settlement was observed to occur after the tailskin had passed completely under the array and the ground began to close in on the tunnel lining before or during grouting operations.
The rates of settlement in advance of the TBM are shown in Fig 4, plotted against depth below ground level. A similar chart showing the rates of settlement behind the tailskin is presented in Fig 5. Boreholes 1A and 5A are located at offsets of 2m from the tunnel centreline and Borehole 2A is located just above the tunnel crown. Close to the tunnel crown, boreholes 1A and 5A are located towards the extremities of the settlement troughs. Therefore, the maximum settlement is less than that at levels directly above and the rate of settlement is seen to decrease at depth.
For Borehole 2A, located directly above the tunnel crown, the rate of settlement increases rapidly close to the tunnel crown. At a level of about 400mm above the crown of the tunnel, the rate of vertical settlement was observed to be 0.06mm/min in advance of the TBM.
Behind the tailskin, the rate of settlement varies between 0.02 and 0.03mm/min at the surface, increasing uniformly down to a distance of about 2m above the tunnel lining. At this point, each instrument is showing a consistent response. Below this level, the extensometers in the boreholes on either side of the centreline show a reduction in the rate of settlement for the same reason as explained above, while over the centreline, at a distance of about 400mm, the rate of vertical settlement increases to 0.14mm/min.
Long-term vertical displacements: precise levelling of the monitoring points in the manholes indicates that, over a period of six months, the magnitude of surface settlements increased by almost 70% from the immediate settlements. Only one further reading has been undertaken since, ten months after the passage of the TBM, indicating that long-term settlements at the surface stabilised after about six months. Fig 6 shows the variation in settlements over time as recorded in Borehole 2A. Each line in the figure represents a horizon, the top one 500mm below ground level and the lowest one 400mm above the tunnel crown. The dashed lines in Fig 6 indicate that precise levels at the surface points were not recorded during this period, so a linear interpolation has been applied.
The same magnitude of settlement over time is not observed at depth. Close to the tunnel crown, a slight heave is evident, possibly due to a redistribution of pore water pressure. Between the two extremes, the increased settlements reduce with depth below ground level. This figure is complicated to a certain extent by the fact that a second phase of grouting was undertaken. The effect of this grouting is more pronounced closer to the tunnel crown, although a permanent upward displacement is evident at all extensometer levels. This upward displacement is not observed when the tunnel first passes through because the continual jacking forward of the TBM would have reduced the pressure of the still plastic grout.
Two observations from Fig 6 are particularly worthy of note: sub-surface heave is still occurring close to the tunnel ten months after the passage of the TBM; and close to the surface, a long-term increase of 70% over the immediate settlements was observed. When evaluating building damage of structures founded at shallow depths, it may be prudent to consider the effect of this long-term increase.
|
“The response of the ground to underground construction like tunnelling is generally the formation of a settlement trough above the tunnel” |
Related Files
Figures 4 and 5
Location map
Figure 3
Settlement diagram
Figure 6
