Following serious failures on two recent major UK tunnelling projects (at Portsmouth and Hull), the British Tunnelling Society (BTS) set up the Closed Face Working Committee to review factual data from causation reports from the two projects. The aim was to identify similarities and differences relating to the two projects and report on any lessons to be learnt or areas of concern worthy of further investigation, and any recommendations for further work.
The failure at Portsmouth (T&TI July 2000, p9) happened after 7.8km of the 8km long, 2.9m i.d. tunnel had been completed. A 15 ring section of lining cracked and deformed, letting water into the tunnel. Deformation of the lining was so severe that it needed replacing. The incident occurred during a shift while the TBM crew were on a break and no work was in progress.
The incident on the 10.5km long, 3.6m i.d. Hull tunnel (T&TI March 2002, p37-41) also occurred after considerable tunnelling was complete, in a section of finished tunnel next to a shaft. Failure was progressive over a length with the tunnel collapsing completely.
Project similarities and differences
Both projects were being constructed with full face Earth Pressure Balance TBMs, using foam/polymer ground conditioning agents. The failures occurred where the ground conditions could be described as “difficult”, with water head in excess of 25m and layered soils, some likely to have no stand-up time. Control of earth pressure balance on both machines was regulated by the driver, with the target EPB pressure for both projects being just above water pressure.
The linings comprised precast concrete, tapered trapezoidal segments with EPDM gaskets. Fine sand was encountered on each drive with grading of the sand resulting in a highly mobile material.
“Volume grouting” was used, meaning that a fixed quantity of grout was available (typically 1.2 x theoretical annulus), and both projects utilised accelerated grouts injected through the rings.
Water in both vicinities was tidal, but not at the tunnel horizon. This was not considered to have influenced either of the collapses, which occurred after a number of kilometres’ of tunnel had been completed.
As for differences, the Hull collapse occurred immediately adjacent to a shaft, whereas the Portsmouth collapse occurred “mid-drive”. The longitudinal performance of the tunnel was considered to be an important factor in the Hull collapse, while at Portsmouth this was not considered to be the case.
At Hull, the section of tunnel at the seat of the collapse had been completed for eight days, whereas the ring at Portsmouth had been completed within the same shift as the collapse occurred. Concrete shear pads were incorporated into the lining at Hull, while the Portsmouth ring utilised polypropylene location dowels on the circle joint between rings. Curved bolts were installed at Portsmouth on both the radial and circumferential joints, while at Hull the ring had temporary spear bolts during erection, which had been removed prior to collapse.
During the investigation of the failure at Portsmouth a quantity of tailskin grease was found in the annulus of the ring, in places the grease had mixed with the grout. At the time of the incident the tailskin greasing system was not working through all ports so an increased volume of grout was pumped through the operating ports to compensate. Peat was present in the crown of the Hull tunnel at the point of collapse.
Risk factors and ground conditions
The following were considered important: ground conditions; TBM operation; and design.
A common feature of the Hull and Portsmouth incidents was the layered nature of the ground, with cohesionless sand layers between cohesive (clayey) layers. The risk with this stratigraphy is the possibility of the cohesionless soils flowing in towards the face, resulting in a loss or loosening of ground in these layers.
Single size fine sand is highly mobile, particularly in combination with significant water velocity, such as leak under high pressure. This material can therefore exploit a very small hole in the lining, possibly as small as a few millimetres and is therefore considered potentially dangerous. There is some evidence that operational vibrations of the TBM increases the mobility of such soil, and it is possible that machine vibrations could lead to densification (and hence a reduction in volume) of loose granular soils. High moisture content organic soils (peat) present the possibility of significant volume change if their moisture content is changed, this could be caused by groundwater leakage into a tunnel.
Recommendation:
If closed face TBMs are being considered for a project it is recommended that site investigations include specific checks for the presence of layered soils, and volumes of fine sand and/or organic soils.
TBM Operation
While it is not considered that EPB pressure was a factor in either collapse, it is noted that no clear guidelines currently exist in the UK with respect to what is considered to be the optimum working pressure for an EPB or Slurry TBM. Normal practice to date has been to ensure that as a minimum, the support pressure is slightly greater than the external ground water head.
With respect to the control of face pressure, it is acknowledged that, even in modern TBM control systems, the principal control is in the hands of the driver.
For an EPB machine to maintain “earth pressure balance”, the driver can vary: the rate of advance; the speed of the screw; opening of the aperture on the guillotine; and the quantity and type of soil conditioning introduced in the head and along the screw. The procedures for control of a slurry shield depend upon the type of slurry shield being used. In pure slurry mode, the face pressure is regulated by the relative speed of the delivery and discharge slurry pumps. Where the slurry is pressurised with a bubble of air (Mixshield type), a much more constant pressure can be maintained.
The present fundamental challenge is clearly to balance the volume of the excavated material with the rate of advance. Several proprietary weighing systems are now available for both Slurry and EPB machines, but the unknown factor is the natural in-situ density of the material (particularly in mixed face conditions). It is considered that in the UK tunnelling industry TBM drivers are generally fully aware of the consequences of over-excavation or loosening of the ground.
Recommendations:
That some research be carried out to examine whether an agreed methodology could or should be established for the calculation of the “required pressure”.
Satisfactory TBM operation and complete grouting of the annulus are clearly required to ensure tunnel stability and both processes currently rely upon the avoidance of human error for success. To date, satisfactory control systems to override human input to these operations have not proved successful. Systematic measurement and recording of data is therefore considered to be the only feasible control at this stage. For the failed section of tunnel at Portsmouth, it is noted that a back-analysis of the data-logger confirmed that the TBM had been operated in accordance with site protocol.
Recommendations:
logging of TBMs. BTS/ICE Specification
currently states that data loggers shall only be
provided if “required by the contract”
Overcut Recommendation:
That some research is carried out into the effect on the ground of the overcut and fluid support (if any), prior to grouting, and that these matters are considered in the design of the tunnel or TBM.
Volume grouting cannot be guaranteed to completely fill the ring annulus and any other voids in the types of ground conditions set out above, particularly where the injection is being carried out through the rings.
From the review of the Portsmouth incident it is likely that the grouting formed part of the cause and effect and to complete the drive a system of additional volume grouting was instigated, which, while slowing production, had the desired effect of no further incidents (albeit it cannot be confirmed that any incidents would have occurred if the original regime had been continued).
In smaller diameter tunnels (up to 3.5m diameter) it is virtually impossible to instigate a pressure grouting system working through the tail without greatly increasing the annulus and it is therefore considered that volume grouting will continue to be used in the smaller tunnels for some time to come.
It should be noted that at Portsmouth extensive investigation along the previously built tunnel concluded that the volume grouting had been effective.
Presently, in the UK, grouting technology still utilises cementitious grouts which can pose problems with pumping and setting times. The grout rheology needs careful thought to prevent excessive penetration into the ground and also to provide adequate early age support of the tunnel ring and internal gantry loads.
Recommendations:
Design
While it is considered that the ring design was not a factor in either of the collapses, two design issues were raised as a consequence of the subsequent investigations. Shear pads were a feature of the Hull ring, and while they assist good ring build, they can potentially lead to damage either during building or in the event that the completed tunnel lining experiences differential movement.
Recommendation:
A review is carried out of international good practice in respect of the use and nature of ring to ring shear systems.
While the ring design is of critical importance, it is also considered necessary to address the longitudinal performance of the tunnel, both in terms of structural design and watertightness. This is particularly the case with potentially critical areas such as entry into or out of shafts, and in zones of soft ground. If the lining is to be bolted or dowelled in the permanent condition, the influence of the bolts or dowels should be considered in the design of the lining.
Risk Assessments
When risk assessments are being carried out, it is recommended that the following potential risk factors be considered: layered soils; fine sand; organic soil; volume grouting; tailskin greasing system; shear pads; dowels; and differential longitudinal movement.
Related Files
Alignment of the 8km long, 3.4m excavation diameter Portsmouth Transfer Tunnel. Marked is the position of the collapse some 200m from Eastney pumping station
