The Dover and Folkestone Wastewater Scheme was initiated by Southern Water to ensure compliance with the EU Bathing Waters and Urban Wastewater Treatment Directives. The Folkestone Interceptor Tunnels Contract is one of the 10 main contracts under the scheme.
The contract has involved the construction of over 2km of tunnels with diameters varying between 1.35m id to 3.66m id. The variety of techniques employed, including two EPB TBM drives, makes the project of particular interest to those involved in tunnelling. All major underground works in the contract are now complete.
Overview
Currently, wastewater from Folkestone is discharged directly into the sea via a short sea outfall at Copt Point. On completion of the present project, untreated wastewater from Folkestone and Dover will be transported to a new wastewater treatment works at Broomfield Bank near Dover and treated wastewater will then be discharged via a marine outfall at Dover.
The contract used ICE 6th Edition. As part of a framework agreement with Southern Water, Mott MacDonald was the designer. Mott MacDonald also carried out site supervision. The contractor was Miller Civil Engineering Ltd (MCEL). It was initially intended that the pumping station works would be under a separate contract, the basic concept being that Folkestone interceptor tunnels would provide the underground space for the pumping station facilities and the pumping station contract would provide everything else required.
However, early in the construction phase, the pumping station contract was also let to MCEL to enable there to be programme benefits and to avoid interface difficulties. The rest of this paper will be devoted to the works in the original contract, ie those related to tunnelling.
As its title suggests, the main function of Folkestone interceptor tunnels is to intercept sewage flows in Folkestone so that they can be carried to the Broomfield Bank wastewater treatment works. In addition, during periods of heavy rainfall, the contract allows transfer of stormwater to a new outfall at the Stade, Folkestone. The two EPB TBM drives were the 1.4km, 2.44m id Interceptor tunnel and the 0.5km, 2.87m id Stade tunnel.
Design
The design process of wastewater tunnels is influenced as much by construction considerations as by the permanent works design criteria. A summary of some design aspects are given below with particular attention being given to show how the construction phase influences permanent design.
At the initial conceptual development stage, the selection of a suitable site for the main pumping station played a critical role. Consideration was given to a number of sites before the disused Folkestone Junction railway yard was finally selected. Its location in relation to the existing sewers and the Stade site was a major factor in its selection. Having identified a suitable main site, the layout of shafts and tunnels could then be established.
Shaft positions were governed by the location of sewers as well as, for Shafts H, E and F in the town centre, the availability of suitable sites. Shaft diameters were determined partly by permanent design considerations and partly by constructability issues. For instance:
The horizontal alignment of the tunnels was determined by the need to connect the shafts and existing sewers as well as constructability considerations. The easiest way to connect two shafts is by a straight line and it is also generally easier to construct straight tunnels than curved ones.
However, the interceptor has curves to avoid the requirement to turn the TBM and all its backup sledges within the confines of a shaft. Also, the Stade tunnel has a curve so that the tunnel aligns conveniently with the secant pile wall at the Stade, which is intersected at right angles.
The interceptor tunnel vertical alignment was defined by the depth of the connections into the pumping station, the required hydraulic gradient and the need to provide clearance beneath the existing sewers in Tontine Street. The Stade Tunnel vertical alignment was defined by the parameters fixed by the design of the pumping stations at the Stade and Folkestone Junction. Both tunnels are approximately 30m deep except for a length of the Interceptor Tunnel near Shaft H, where the cover reduces to only 6m.
Clearly, while permanent design considerations define the minimum diameters of the tunnels, the actual tunnel diameters were determined by other factors such as the availability of precast pipes,TBMs and segments and the space required for construction. For example, the Shaft 1 to Shaft 5 tunnel is of a sufficiently large internal diameter (3.66m) to allow the Stade TBM, having been assembled and prepared in Shaft 1, to be shoved through the tunnel and launched from Shaft 5, thus permitting other activities to commence in Shaft 1. Availability of the Lovat machines prompted the contractor to request changes from the tender design, which the client accepted, to the tunnel diameters specified on the tender drawings which showed 2.1m id for the interceptor and 3m id for the Stade.
The defining criteria for the depths of tunnels was generally the topology of the routes and the positions of the connecting structures, not the ground conditions although these were borne in mind during the design phase. From the Folkestone Junction site to Shaft H, the ground surface level is falling. From Shaft H to Shaft F, the ground surface level rises steeply. Elsewhere, i.e. within the main site, towards Neason Way and the Stade, the ground surface is relatively flat.
The geological sequence is as follows. Made ground overlying Gault Clay which in turn overlies Folkestone Beds which overly the Sandgate Beds. Gault Clay is a stiff grey clay. Folkestone Beds are a medium grained sand with isolated rock lenses called doggers.
The presence of these doggers added a different dimension to all excavations through this material. Sandgate Beds are a silty fine sand. The water table is approximately 12m below ground level. With the exception of the Warren Road pipejack, which was in Gault Clay, all the tunnels were constructed in either the Folkestone Beds or the Sandgate Beds.
Construction
This section will describe the construction of:
Shaft 1 and Shaft 4 were constructed using diaphragm wall techniques. Shaft F was constructed using the jacking technique and all the other shafts, with the exception of Shaft E, were underpinned. Shaft sinking activities have generally been carried out using standard techniques which worked well.
Site organisation and initial planning
It was always envisaged that the Folkestone Junction site would be the main site for the contract and also for the pumping station works. Site preparation was carried out under an advanced works contract which involved removal of contaminated land and construction of temporary access roads. This enabled the site to be handed over to MCEL in an uncontaminated state, thus eliminating any possible risks associated with having clean and contaminated zones on site.
The site at Folkestone Junction was constrained in that a lot of separate activities needed to be carried out concurrently. Shaft 1 was particularly critical in that pumping station construction could not proceed until tunnelling works were either completed or transferred elsewhere, ie to Shafts 5 and 4. Careful thought to programming was required to allow activities to be carried out concurrently and thereby take full advantage of the benefits of placing both the tunnelling and pumping station contracts with the same contractor.
Activities were carried out from and around Shaft 1 as follows:
The Warren Road pipejack tunnel was carried out from Shaft 3 so the timing of this activity was not particularly critical. Shaft sinking in the town centre was also not critical.
Tunnelling operations were generally carried out 24 hours a day, five days a week. The facility for night shift working was useful, particularly early on, to allow different activities to proceed concurrently.
Hand drives
The two 1.83m id inlet and overflow tunnels and the 3.66m id Shaft 5 to Shaft 1 tunnel were constructed using hand shields equipped with a tunnel lining erector. All three tunnels were in the Folkestone Beds and the doggers which were encountered had an influence on progress. Stitch drilling, breaking by use of drills and feathers and various pneumatic tools were used to remove the rock.
The 1.83m id tunnels were both driven from Shaft 1 to an existing sewer that runs deep under Morrison Road and Folly Road just outside the site boundary. The inlet tunnel will intercept this existing sewer which currently carries 70% of the wastewater from Folkestone to the existing outfall at Copt Point. Conveniently, these two tunnels were shallow enough for the drives to be completely within the influence of the dewatering system on the Folkestone Junction site.
The dry ground conditions, the compacted nature of the sand, the small diameter and, to an extent, the presence of doggers meant that the face was generally stable. Typically, there was a thick, horizontal band of rock at the centre of the face. This band of rock varied in thickness such that it sometimes extended into the crown. Tunnelling was stopped just short of the existing sewers and the connections made later.
The contractor used an inventive way of tunnelling both drives whilst Shaft 1 was excavated below; Shaft 1 was of a sufficient diameter to allow there to be a platform outside each tunnel from which tunnelling could continue.
The 3.66m id tunnel (the largest on the whole project) was driven from Shaft 5 to Shaft 1. The tunnel was some 30m below ground level, which meant that it was just below the dewatered level of the water table. The pattern of dogger formation was generally: a central horizontal band of almost constant thickness and stronger material in the invert which varied in thickness at times as far as the central band. Water seepage occurred above axis level and in between the bands of rock because the tunnel was just below the influence of the lowered water table.
The invert rock was difficult to remove using the methods described above and progress was restricted to an average of two rings/12h shift. Stability of the face was maintained using face boards above axis level; generally the presence of rock made boarding unnecessary below invert level. The central band of rock acted as a support for the timbers. Entry into Shaft 1 involved breaking through the 1m thick diaphragm wall.
The 3.66 id drive gave a valuable early indication of the material to be encountered on the TBM drives.
Warren Road pipejack tunnel
The 200m long 1.35m id tunnel was constructed from Shaft 3 through Gault clay, which is a stiff grey material, using standard pipejacking techniques. Excavation was carried out using a backactor which was manually operated at the tunnel face. Bentonite injection was carried out for the latter half of the drive only, as the tunnel here had greater depth and the clay was stiffer.
TBM drives
Separate Lovat earth pressure balance (EPB) TBMs were used for each of the Interceptor and Stade drives. The principle of an EPB TBM is that the earth/water pressure acting on the tunnel face is resisted by a plug of spoil inside the plenum chamber. As excavation proceeds, muck enters the plenum chamber and is taken to the muck belt via a screw. Removal of muck from the plenum chamber is controlled in order to maintain the required pressure to resist the forces acting on the face.
The choice of TBM was influenced largely by the position of the water table and the medium sand/silty fine sand ground conditions. The cutterhead pick tools were mounted on Trimec titanium alloy backing plate of a similar type used on the Fylde Coastal Water Improvement Scheme tunnels where boulders were encountered. Both interceptor and Stade machines were equipped with flood doors at the head which could be closed to avoid flooding the tunnel in the event of system failure.
Particular attention was paid to grouting. The grout mixture contained an accelerator to enable it to achieve a fast initial set. This initial set could be reversed if the grout pressure increased, thus enabling all voids behind the ring to be filled. Grout pressures, accelerator quantities and grout intakes were monitored during the operation to provide reassurance that the EPB system was controlling ground movement effectively. The effect this had on settlement is discussed later.
Before moving on to discuss each drive specifically, it is helpful to briefly state the sequence of events.
The Interceptor drive commenced first from Shaft 1 to Shaft 4a. At Shaft 4a, the condition of the cutterhead tools was assessed and appropriate modifications were made to both the interceptor and Stade cutterheads. The drive shaft was transferred to Shaft 4a and mining proceeded until a mechanical breakdown occurred some 40 rings from Shaft 4a.
The Stade drive commenced from Shaft 5 and proceeded until ring 300, when it was necessary to repair the cutterhead tools. Access to the head was facilitated by a timber box heading being constructed around the machine from an opening made in the skin of the machine. Details of the box heading are given in below. Almost concurrently, access to the cutterhead of the Interceptor machine was gained also using a similar box heading. The Stade drive was completed. The interceptor drive was recommenced soon after breakthrough at the Stade.
Stade Drive
Any discussion on the Stade drive will be dominated by the effect of the doggers. Experience gained from the 3.66m id hand drive showed that encountering hard dogger material on the Stade drive was extremely likely. Start of this drive was delayed until the condition of the interceptor TBM cutterhead was assessed, so that the same modifications could be made to both machines.
The modifications involved changes to the basic dimensions of each pick and to the configuration of picks on the cutterhead, such as to increase the tendency for energy to be used to break up the rock and decrease the tendency for energy to be used in wearing down the picks. Despite these modifications, the cutterhead was still only suitable for use in soft ground as it was hoped that the length of drive in rock bearing material would not be too long.
As soon as the Stade drive commenced, local residents reported a loud rumbling sound in their homes. Site staff who visited the local residents verified that this was true. Although it seemed difficult to believe that the machine (which was 30m below ground) could be heard and felt at ground surface, local residents could give exact times for when the machine was shoving based on the time of the rumbling sounds. The sound was caused by the TBM mining the rock.
Failure of the knuckle joint inside the screw conveyor occurred often, probably due to the passage of lumps of rock through the screw. EPB mode was not used for most of the drive because the quantities of rock encountered meant that ground conditions were unsuitable for the formation of a plug in the plenum chamber. While rock continued to be encountered, as indicated by the contents of the muck belt, the average progress rate was around five rings/shift.
As the TBM approached ring 300, the shove times for 1 ring (1m) increased from 50 minutes to three hours. The contractor decided to stop the machine at a convenient location least sensitive to settlement ie under a pedestrian footpath, and proceed with repairs to the head.
Access to the cutter head was facilitated by a box heading constructed around the TBM. The water table was initially above axis level of the TBM. Had the TBM been completely beneath the water table, the operation of driving the box heading would have been too hazardous to execute in free air and the application of LP compressed air would have become a necessity.
However, the contractor drilled two small drain holes through the TBM tailskin below axis level to ensure the water table was reduced locally prior to cutting through the skin and commencing the box heading.
At the same time the screw conveyor was disconnected and ground water was allowed to “bleed” from the face through the TBM cutter head and plenum chamber. Once the contractor was certain that the water table around the TBM was reduced, and that sand wash out would not occur, the tailskin of the TBM was cut into and the heading commenced.
A section of the tail skin had a 1m x 1m cut-out already installed above axis level and the cover over this was removed to allow the close boarded box heading to be driven out at right angles to the TBM. After advancing outwards for 2m a specially fabricated steel roof support was built in the heading to enable the driveline to be turned through 900 and then to be continued towards the front of the machine.
The steel “table” was used to ensure that the ground was fully supported all round at the change of direction, the steel legs of the table both supporting the crown of the heading as well as the two close boarded sides of the four sides of the table, the other two being open and at right angles to each other.
The heading was then continued parallel to the left hand side of the TBM leaving a 1m wide pillar of undisturbed ground between the two. It also had an upward slope to ensure that the inbye end of the heading crown emerged some 200mm above that of the TBM cutting head. The heading was advanced in sets of 225mm x 75mm timber placed contiguously then grouted up in pairs with all joints well sealed before grouting. This method of working ensured that the ground was fully supported each 450mm of advance.
The ground at this point in the tunnel was still in the Folkestone beds so the heading encountered extensive rock throughout its length, which all had to be broken and removed by hand working. This, together with the cramped conditions, necessitated continuous rotation of the men at the face.
When the heading had reached a point some 2m beyond the head of the TBM, it was again turned through 900 using the technique described above, and driven back across part of the face of the machine. The heading was then opened out into a small working chamber which exposed a quadrant of the cutting head, the side boards on the right being supported off the extrados of the TBM, while those to the left side were founded on foot blocks and undisturbed rock. Additional support to the timber chamber was also provided by use of RMD soldiers.
By carefully cranking round the TBM cutting head through the exposed quadrant the entire head was progressively inspected and refurbished, with every pick changed.
When the TBM remedials were completed the chamber was backfilled using both loose pea shingle and sandbags filled with pea shingle. As various support members of the chamber intruded upon the drive line of the TBM when it restarted, parts of the chamber had to be dismantled in the reverse order in which it was mined, the sand bags being inserted as disposable replacement supports for each timber and steel member removed.
Once the chamber backfilling was completed the timber heading was backfilled, with grout pipes built in for back grouting. The TBM tailskin was reinstated and the heading was back grouted. The TBM was restarted and when it had progressed beyond the immediate area backgrouting through the rings was again undertaken to ensure that any possible voids were filled.
Various precautions were taken to avoid excessive surface settlement in the vicinity of the box heading. These included:
This whole scheme to refurbish the TBM head was designed by the contractor’s head office temporary works department and the method statement was developed by an iterative process between the site and head office teams. The whole rescue operation, including design, took just three weeks of continuous working to achieve and no significant surface settlement occurred.
After this enforced stoppage, progress was around 12 rings/shift. However, this good progress rate was not solely due to the effect of the refurbished head; at around ring 360, the TBM began to pass through soft, non-rock bearing material which appeared to be a transition zone between Folkestone and Sandgate Beds.
The machine finally broke through the secant pile wall at the Stade site without incident.
Interceptor drive
A box heading, similar to the one used on the Stade, and as described above, was also carried out on the interceptor drive to gain access to the head for repairs. This time, the machine was just outside the site boundary and still within the influence of the Folkestone Junction site dewatering system so no drainholes were necessary.
Repairs were carried out on the machine and the drive was recommenced. However, this time rock continued to be encountered and the progress rate continued to be poor – caused in part by shove times of around three hours. It was clear that the cutterhead was suffering excessive wear and required refurbishment. As the TBM was now below the water table, it would have been too hazardous to construct another box heading in free air, so the contractor decided to install facilities for using compressed air. This involved plant procurement, provisions for bulkhead doors as well as medicals for relevant site staff.
Paradoxically, the facilities for compressed air working were installed but never used because soon after the installation was completed, the progress rate increased, reaching up to 15 rings/shift. This was because the TBM started to pass through non-rock bearing strata. The TBM reached Shaft H without further difficulty and the head was refurbished there before the drive was continued. The TBM reached its destination, Shaft F, with no further incident.
Settlement
Generally, the 30m depth of the tunnels meant that the maximum predicted settlement was less than 5mm for most of the drive. Actual measured surface settlement was 0mm-3mm for most of the drive, ie negligible. The control over the grouting system was no doubt a contribution to this.
Surface settlement monitoring consisted of two methods; namely a relatively unsophisticated system involving daily levelling of Hilti nails installed in the road and structures as well as deep level monitoring stations. The deep level monitoring stations, installed along the approach to Shaft H, were used to measure subsurface settlement accurately and hence face loss. Thus, the ability of the machine and grout system to control ground movement was confirmed prior to the machine reaching the area of low cover around Shaft H, which was always perceived as being the most at risk from excessive settlement.
In the vicinity of Shaft H, the tunnel passed below two old brick lined sewers with under 2m clearance. Prior to tunnelling, the contractor carried out compensation grouting, installed soil dowels around the sewers and installed electrolevels inside both sewers. These electrolevels were linked to a PC and continuously monitored whilst the sewers were undertunnelled.
Mainly due to these protective measures, the TBM passed through the low-cover zone without any unforeseen incidents.
Conclusions
The Folkestone interceptor tunnels has been notable because of the variety of both the tunnelling methods used and the ground conditions encountered. Tunnelling by EPB TBMs, by traditional hand mining and by pipejacking were all carried out at this one site and within one year.
Despite the difficulties arising from the sandstone doggers the TBM drives were completed within the time allowed for on the programme and the tunnelled aspects of the contract are now complete.
Related Files
Layout of tunnels
Folkestone Junction Pumping Station
Stade TBM
