The Briars Lane area of Hatfield is close to the town centre and consists of housing, playing fields, and a school. Upper Chalk lies approximately 5 to 10m below the surface, and it is this that attracted miners in the 19th and early 20th Century. In the process of extracting the rock, a network of shafts and narrow adits were formed. Chalk mining ceased around 1920, with most mines abandoned, and very few records left.

Hatfield, some 20 miles north of London, expanded in the mid 20th Century, and from 1937 the Briars Lane area was developed for housing. No attention was paid to the mines beneath until, in 1978, the first major subsidence event occurred with a plan area of 4.6m by 3.6m. A second event in 2001 resulted in a 6m deep hole with the collapse of a footpath and some property damage.

Project Manager Mace and Designer and Contract Supervisor Hyder were brought in by Welwyn Hatfield Borough Council in Partnership with Hertfordshire County Council in 2005. The resulting risk assessment required the evacuation of a number of at risk properties and the closure of the school. This suddenly gave the project a very high public and political profile, and the team had to move rapidly.

Scheme development

The scheme objectives were to allow safe occupancy of the site, to ensure minimal future movement from the mine workings, and to restore public confidence. At the same time the works had to be carried out with a minimum of disruption to residents, and within budget and programme.

Between March and July 2005 a ground investigation programme was carried out, which defined the core site as an area of around 3 hectares. A feasibility study from September to December defined the scope of the proposed works, and a risk assessment in November 2005 resulted in the evacuation of a number of properties. In December 2005, a successful application for funding from English Partnerships was made. Site works were carried out between August 2007 and March 2008, and the properties re-occupied in June 2008.

A desk study and public consultation process revealed a history of ground movement dating back to the original development, and evidence of a complex network of galleries beneath the site. The chalk had been extracted, fired to form lime and then spread on fields to reduce acidity and promote growth, and an old map of the area showed a kiln present. Certain horizons of chalk are especially pure and it is likely that the miners followed these, identified by colour and softness.

In the development of a solution, there were a number of key risks which included the efficacy of the treatment, the effects on the adjacent properties, maintaining stakeholder and public confidence, and the unknown numbers, locations, and volumes of the voids. In an enlightened move, the Client retained responsibility for these risks, allowing the integrated project team to manage these whilst concentrating on developing and implementing the best technical solution. A joint risk register was produced and maintained throughout the project, and this placed significant emphasis on obtaining high quality data, controls during the implementation phase, and a comprehensive validation exercise.

Site investigation

The first challenge for the team was to ascertain the extent of the workings and their volume. 82 boreholes were drilled, and surveyed using down-the-hole laser, microgravity and CCTV. A further 1581 dynamic probes were driven to typical depths of 20m providing a record of penetration resistance against depth.

Investigations showed that there were up to 4 levels of workings between 6m and 20m deep, each typically 2m wide and 2 to 3.5m high, but these were of varying width and irregular orientation, and were often partially or wholly blocked from collapsed material, leaving disturbed ground above.

Procurement

An Engineer’s Design was developed employing foam concrete backfill of workings and compaction grouting of voids. It was appreciated that the need for design modification and treatment methodology refinement would be necessary during the treatment works to respond to variations in ground conditions and the extent of the workings encountered. Consequently a high quality data management system was demanded of Contractors to allow real time monitoring of the efficacy of the treatment works and extent of treatment. In order to allow realistic validation of the treatment a range validation criteria were set including additional dynamic probing.

The emphasis on quality was highlighted to prospective Contractors throughout the tender process, and tenderers were marked on a 90% quality assessment of technical and risk management abilities, with integrated working within a true partnership key, and only 10% on price. This led to the appointment of BAM Ritchies as Main Contractor on an NEC Option B Contract, with RJM Ground Solutions as their technical advisors.

Public interface

Keeping the public informed and minimising disturbance were key scheme objectives. Regular meetings with residents groups were arranged, and information was passed on through regular letter drops. Sites were erected and dismantled daily, and the batching plant was provided with shielding.

Building condition surveys were carried out on all properties and an extensive monitoring system installed, consisting of continuous remote monitored tilt meters and precise levelling surveying during treatment works. Data was continuously logged and had trigger levels been exceeded, texts would be sent to key personnel. This system was tested, but no trigger values were exceeded during the works.

Void filling

Some 675 vertical and 120 inclined holes were drilled for the bulk filling exercise, typically 20m and 35m long respectively. A variety of rigs were used including small manoeuvrable tracked rigs, and larger excavator mounted drilling booms for reaching over hedges and fences, always aiming to minimise disturbance. The project included a contingency for drilling from within houses and the necessary reinstatement, but the need for this was successfully avoided using inclined drilling.

All holes were surveyed using a standard drainage CCTV system, and those indicating voids were re-surveyed using a directional CCTV, which helped to define the void shape and location, though volumes were very difficult to estimate. All holes that had intersected substantial voids were then filled with a foamed concrete with a density of 0.5t/m3. A total of 2240m3 was injected, with two holes taking 300 and 350m3 each.

Compaction grouting

With the larger voids filled, a comprehensive exercise of compaction grouting was employed to fill any remaining voids and consolidate and strengthen zones of disturbed chalk. There was a nearby drinking water extraction point from the chalk aquifer, and the Environment Agency would not allow the use of PFA within the grout mix. To maintain pumpability, a replacement bulk filler was required, and the team successfully carried out off-site trials of a low slum mortar mix using limestone dust. However, the short material life of this mix combined with the high head losses experienced during initial site use demonstrated the need for a more pumpable thinner grout with a longer life, and the team quickly developed this. The final 8:1 mix (limestone dust:cement) developed approximately 1 to 2N/mm2 at 28 days, and a total of 1920m3 was injected during the project.

Inflatable packers were employed to inject in typically 2m stages in each hole, with injection pressures of 10 bar targeted below 10m reducing to 5 bar between depths of 5-10m. Injection pressure was measured immediately adjacent to the top of the hole, along with an accurate volume measurement that was subsequently reconciled against batching records. The team suspected that two early instances of packer failure were due to high injection pressures forcing grout around the packer. Subtle modifications to the injection procedure and the switch in mix design prevented reoccurrence. To ensure maximum plant mobility and minimum disturbance, the injection pumps were fitted to site dumpers, running from the vehicle hydraulics. As part of the validation process a further 251 dynamic probes were undertaken.

Validation

From the outset, it was clear that the management of data was a major exercise with over 375,000 dynamic probe data points along with dozens of ground investigation logs and many hours of CCTV footage. To deal with this challenge the Rockworks 3D mine modelling software package was selected, and the use of this from the outset allowed the team to address the inevitable uncertainty by permitting (figure1): Continuous review of progress; real time discussion on the efficacy of the treatment and the need for additional work; sectional validation of the works (and hence site re-occupancy).

Dynamic probe data was entered into the model with greater than 5 blows defined as intact chalk, 3-4 blows as disturbed chalk, and 0-2 blows as a void. Figure 1 shows the mine model developed from the pre-treatment data with voids and disturbed chalk shown red and intact chalk shown blue. Bulk filling volumes and grouting volumes per injection stage (figure 2) were fed into the model as these took place, resulting in a continuously updated model of the remaining disturbed ground. This allowed the team to assess the treatment carried out so far on a weekly basis, and to target the next stage to maximum effect. The Rockworks output also formed a key part of the multi point validation process allowing areas requiring more treatment to be identified. A few locations required minor retreatment, but the majority showed that first stage filling had been totally successful.

Validation reports were produced summarising data from drilling, void filling, grouting, and the validation probing exercises to demonstrate that the residual risk of voids was extremely low. To enable early re-occupancy of site, a total of 28 sectional reports were produced.

Conclusions

The team put the success of the project down to the early acceptance that change is inevitable and must be managed by the whole team working together in a framework of constructive cooperation. Real-time data management enabled the assessment of treatment efficacy, allowed informed modifications to be made to the treatment methodology and permitted validation to be carried out during the treatment process. Above all, mine treatment schemes like Briar Lane are about confidence, and this can only be achieved by a high quality product. T&T


Questions from the floor

Alan Webster of the IMMM asked what level of monitoring is continuing following project completion.
Monitoring of the treated area forms a part of an on-going wider study for the Hatfield area (which includes items such as emergency planning), and the team are very confident that the residual risk from further ground movement in the treated area is negligible.
Dr Ponnampalam Linkeshwara of Network Rail asked how many properties the work covered and what the cost of the works were per property protected.
The project covered approximately 100 houses and with a total budget of US$5.8M, the cost per property was therefore around US$58,444.
In a supplementary question, Dr Linkeshwara asked what lessons were learnt and what would be done differently were the project to be repeated.
Two main points were raised: The lack of confidence in attaining the required fill using a stiff mortar grout mix as recommended by the CIRIA grouting guidance and the uncertainty in assessing this pressure at the point of injection when measured at the ground surface when high head losses are occurring along the injection line.
David Hartwell, Consultant asked whether there was any evidence of heave or local fracturing into permeable zones.
The injection pressures were chosen so as to avoid ground heave, and the upper 5m of ground were not treated except where open voids were noted. Due to these controls, no significant heave was noted. Two injection packers were lost during initial grouting works, but the team consider that this was due to high injection pressures causing ground movements and grout flow around the packers. This did not re-occur following the change in grout mix and injection procedure.

Fig 1 – 3D model from dynamic probe data Fig 2 – Compaction grouting data Batching plant shielding Dynamic probe rigs in action Vertical bulk filling holes being drilled