Investigations into the collapse of the A2 in south east London took almost four months and revealed a widespread problem of backfilled chalk mines.
Paul Wheeler reports.
It is a quirk of the UK financial calendar that most new authorities and trusts are inaugurated on 1 April - April Fool's Day.
But for Transport for London (TfL), its first week as an allembracing authority with a remit to keep the capital moving, was anything but a joke.
On Sunday 7 April, a 6m-wide crater opened up on the Blackheath Hill section of the A2, the main arterial route into London from the south east. The road was immediately closed, diversions set in place and 37 households evacuated (GE June 2002).
'For the first few days we didn't dare get any heavy equipment on site, ' says Trevor Williams, TfL Street Management's southeast area manager of roads. 'All we could do was carry out levelling up and down the road to detect movements.
'There was total ground loss below a section of the road and adjacent pavement. It was clear the failure was gradually enlarging as the collapsed faces assumed their angle of repose.' TfL's priority was to stop the hole getting bigger, so it filled the void with about 100t of pea gravel - quite a challenge when it was unsure of the stability of the ground beneath workers' feet.
Within a week TfL, term contractor Fitzpatrick and consultant Parkman were confident the collapse had stabilised. With the engineering effort focused on getting the road reopened, the team began a very large site investigation, covering a 600m section of the A2 where it drops down from Blackheath to Deptford.
When the collapse occurred, the assumption was that the road had fallen into a natural chalk solution feature or ancient chalk mine workings.
The desk study revealed a very different situation. It seems the A2 lies precisely over the alignment of a Roman road. From the Middle Ages the chalk underlying the ground surface either side of the road was quarried and burnt in kilns to make lime.
Quarrying activity is marked on maps dating from 1695 to 1870, with the latter showing the full 300m by 250m extent of the quarrying to the south of the road.
This is probably why there are no Victorian houses in the area.
Since then the quarries have been backfilled and the area developed with housing. The road has been widened so that its shoulders now extend over the edge of the mine workings.
The main strategy for site investigations was to use a geophysical survey backed up with boreholes to discover the extent of the quarrying and the nature of the backfill.
'We wanted to use two techniques so that one would be proving the other, ' explains Williams.
For the geophysics, Parkman called on the services of Keele University-based Microsearch, which carried out a microgravity survey along the entire section of concern.
The basic principal was to position boreholes where the geophysics identified anomalies, but there was some concern over the weight of drilling equipment and the fact that most site investigation techniques use percussive drilling methods. An early priority was finding a company that could offer lightweight window sampling with a rotary coring follow-on.
Bristol-based Structural Soils had the right equipment and was able to start immediately. The project subsequently developed into a 140-borehole investigation.
Boreholes were up to 45m deep, drilled using up to three rotary rigs, a dynamic probe rig and two cable percussion rigs. About a quarter of the holes were cored while the rest were open-holed.
Structural Soils needed to subcontract more rotary drilling crews, but they proved to be in surprisingly short supply. After a nationwide search three were eventually tracked down and recruited.
Investigating the full subsurface extent of the collapse feature presented some additional problems, mainly because of concerns over the area's stability. Williams says TfL was also keen to maintain the two-technique approach to site investigation, so that one method proved the findings of the other.
Parkman enlisted the help of Cementation Foundations Skanska to use its '3dT' borehole seismic tomography techniques to produce a 3D image of the ground (GE December 2001). Four boreholes were drilled around the perimeter of the collapse, down which Cementation placed seismic source and receiver arrays.
By sending and receiving seismic signals between the boreholes it was possible to build up a 3D velocity model of the ground.
The resulting velocity contour image can be viewed on a computer from many different angles and gives clear definitions of the contrasting densities of different structures in the ground.
According to Cementation, 3dT clearly defined the collapse feature, as well as the deeper underlying ground structure.
Parkman was initially concerned about the quality of seismic signals travelling through the dry granular fill material, but after a series of trials using a variety of down-hole tools, the team was very pleased with the quality of the data.
All this work revealed that quarrying had left a thin, 20m high intact chalk 'spine' running along the line of the Roman road, overlain by a few metres of dense Thanet Sand.
This means the edge of the road and surrounding areas are sitting on a largely granular quarry backfill - typically mixed sand and chalk fragments but with the occasional brick giving away its man-made origin.
The collapse has been attributed to ground loss a result of water removing fines from the underlying granular fill material.
The water source - whether it was natural flow within the ground or from a leaking water main - will probably never be determined.
Two water mains do lie within the collapse zone but, as Williams points out, when water mains rupture, soil fines are usually washed up onto roads. In this case everything went down the hole, suggesting the mains broke during the collapse.
There is a history of recent movement in the area, evidence of which includes two abandoned blocks of flats and a derelict pub.
Within the zone of the quarrying, the road is - even before remedial work and notwithstanding the collapse - arguably the safest area.
The local authorities in whose area the collapse occurred, Greenwich and Lewisham borough councils, are extending the ground investigation to establish the implications of the collapse.
These instances of structural movement can no longer be assumed to be associated with isolated chalk solution features.
They could instead be the legacy of the quarrying, an activity that persisted for hundreds of years but appears to have been forgotten less than a century after it stopped.
From the microgravity and borehole work, TfL identified a 180m section of the road that needed to be treated, by drilling and grouting, before it can be reopened to traffic.
Fitzpatrick brought in contractor Keller Ground Engineering to carry out remediation (see box).
Williams believes TfL's strategy of investigating the cause fully is one that will prove to be the most cost- and time-effective in the long run.
In the meantime he is highly complimentary of the 'partnership approach' that has seen everyone - from the police, local authorities, residents' associations, Thames Water, the Environment Agency and even the bus companies - making the best of bad of situation.
'We've requested that drivers avoid the area and we believe that many have changed their travel arrangements, ' he adds.
Williams is also grateful for the co-operation of Greenwich and Lewisham borough councils, which have relaxed some local traffic restrictions to enable motorists to bypass the area more easily, although all have remained conscious of residents' needs and avoided opening up rat-runs.
In response to local concerns TfL has tried to maintain corridors for local businesses; bus services have been rescheduled and for a while TfL operated a shuttle bus between the two ends of the restricted area.
'We appreciate it's not a good situation, ' says Williams, 'but people understood the limited options available.'
Keller Ground Engineering started grouting in late July. It is a two-phase process and if carried out correctly the injection works are self-checking.
In the first phase Keller drills holes on a 3m grid through the affected ground, penetrating the insitu chalk by 2m. Grout is injected into each hole until the ground will not take any more. After the grout has gone off, a secondary grid of injection holes is drilled in between the primary grid. From the grout take in the secondary injection holes it is possible to check the effectiveness of the grouting and assess the performance of the works.
Groundwater monitoring wells and 20 inclinometers were installed to help monitor ground movements during grouting. Instrumentation firm ITM, working for Keller, was originally manually monitoring these five times a day but to speed up operations it is installing six strings of in-place inclinometer sensors to give real-time ground movement data. As grouting proceeds along the road the sensors will be moved to keep track of work.
The hope is that the spacing of the grout injection holes can be increased to reduce the amount of drilling needed to cover the treatment area.