Grout bags and tubes-à-manchette pipes have been put to novel use deep beneath a decommissioned oil refinery on the Thames estuary in Essex.
They were key tools in geotechnical contractor Bachy Soletanche's innovative approach to sealing a redundant cooling water tunnel which posed a flood risk in the redevelopment of the site.
At ground level was little to see except a drill rig sinking boreholes on the banks of the estuary near Southend.
But 19m below, in the dark confines of the 50year-old tunnel, a technical battle was being waged.
Strong tidal forces, soft silts and hard cast iron linings were among the challenges facing the geotechnical team.
Work to seal and backfill a 110m length of the tunnel and its integral 25m deep shaft was finished last month, two weeks early.
The brief from client Shell for the £300,000, 10week contract was simple. The 8km 2Shell Haven refinery is being demolished to make way for industrial redevelopment and its 3m diameter cooling water inlet tunnel had to be sealed.
The 285m long tunnel, with its outlet 168m into the estuary, ran inland beneath the 3m high concrete wall overlying 8m deep sheet piling which form the riverbank's flood defences. The tunnel fed into a deep shaft within which cooling water rose a few metres into the adjacent pump house.
But in the redeveloped complex, the tunnel offers a direct route past the flood defences - a possibility unacceptable to the Environment Agency. The tunnel had to be sealed where it passed beneath the shoreline and backfilled along the 110m landward section and the shaft.
The obvious solution was to weld a steel bulkhead into the tunnel and, through this plug, concrete the landward end. But Shell was concerned at the thought of putting men in such confined spaces, let alone divers with welding torches. Bachy Soletanche was called in to suggest a remote solution, working only from above ground.
'Safety was more important than cost, though our option was still marginally cheaper, ' says Bachy Soletanche operations manager Alec Courts. 'Our idea was novel though simple and made use of standard geotechnical techniques. '
Nonetheless the operation was fraught with unknowns, risk and - in the event - suspense when the initial tunnel plug suddenly failed.
The plan involved sealing the tunnel with two large grout-filled bags. The 7m gap between them was to be concreted to create a plug before the rest of the landward tunnel and its shaft were backfilled with a mixture of new concrete, crushed concrete aggregate, gravel and grout.
Challenges began before work even started.
'We did not know the tunnel's exact location, condition, lining details or ground makeup above it, ' Courts recalls. 'Shell and the Environment Agency were requesting detailed method statements from the outset, but we needed to remain flexible to adapt to what we found. '
Exploration boreholes revealed the tunnel to be formed of 20mm thick cast iron segments with an internal 200mm thick secondary concrete lining, and half full of river silt.
The seal was achieved by drilling two 178mm diameter boreholes into the tunnel beneath the riverbank and 7m apart. A rolled-up 3m long geotextile grout bag, with a tube-à-manchette pipe inside it, was lowered down each hole.
Previously the contractor had used only small diameter grout bags as borehole sealers and these larger cylindrical versions had to be specially manufactured. Once positioned, each bag was inflated with air and then filled by pumping down a total 8m 3of grout dispersed by the tube-à-manchette.
Nozzles in the pipe were controlled to allow two days of gradual filling from the bottom, enabling the bag to adopt the exact tunnel profile.
Close monitoring of pumping pressures gave a clear understanding on the surface of what was happening in the tunnel where a tight grout specification was crucial.
The 14t of grout injected slowly into each bag needed to remain fluid for at least six hours while the sides expanded to fill the tunnel. The aim was to encourage free water to escape through the bag's geotextile lining as the grout formed a gel on its inner surface.
This gave it rigidity to profile the tunnel shape before the bag's contents set solid. A 0. 45 water/cement ratio and seven-day 40N strength was specified.
Water and silt were sucked out from the 7m gap between the bags before the space was filled with a flowable 35N concrete mix pumped down additional boreholes. The grout bags provided solid stop ends and the 48m 3concrete plug, poured in one day, cured sufficiently overnight to create what appeared to be an effective water seal.
But before backfilling the remaining tunnel length with concrete, water and silt had to be pumped out using the linked shaft.
When the tunnel was open to the estuary, the river's 5m tidal range had been mirrored in the shaft with water levels rising and falling daily.
During four days of pumping to empty the tunnel, the gradually subsiding water level in the shaft fell at a constant rate, oblivious to tidal changes, and proved to engineers that the seal was sound.
'But mid-morning on the fifth day, the water rose 9m up the shaft in seconds, ' Bachy Soletanche contracts manager Simon Young recalls vividly.
'Our instant suspicions that the concrete plug had blown were soon confirmed when we sank new boreholes into it. '
These revealed a 500mm deep void extending the full length of the plug directly beneath the tunnel crown, an area that should have been solid concrete. It is now assumed that the jetting and sucking out of the 1. 5m layer of silt in the tunnel invert between grout bags had been only partially successful.
A thick layer of silt had been left in the invert and concrete then poured over it. This had created a weak zone and later, the combined pressure on the plug of an incoming tide on one side and water being pumped from the tunnel on the other forced this silt out and past the inward grout bag.
It is thought the concrete plug then moved downward, creating the void in the crown and providing a route for tidal water to flood through the seal and into the tunnel. The solution was to refill the void with concrete and strengthen the inward grout bag with a concrete wedge poured against its landward face.
The rest of the tunnel has now been backfilled with 350m 3of a weak mix C20 concrete. This was preferred over grout to avoid the possibility of seepage through gaps between tunnel lining segments.
The shaft was filled with gravel and crushed concrete offered free from the nearby refinery demolition. Finally, the 4m thick gravel zone between the main tunnel inlet and its higher outlet into the pump station was strengthened by grout pumped through tubes-à-manchette pipes, this time being used in their more conventional role.
'Innovative and safe' is how Shell's senior civil engineer David Wayne sums up the chosen solution. 'At the start we really had no idea what we faced down there, but we knew we did not want to put men in the tunnel.
'This idea has worked so well we are likely to consider a similar solution to a leaking water tunnel in our refinery at Stanlow Cheshire. '
But the most unusual specification at Shell Haven had little to do with ground engineering and a lot to do with concerns about terrorism.
Every ready-mixed concrete truck arriving at the refinery gates had to be escorted by a nominated site worker to the tunnel site. Even a decommissioned oil refinery is, it seems, a risk too far.