lEarly involvement of geotechnical engineers was key to success of a housing development on former marshland in the Thames estuary. Shon Williams and Stephen Durham report.
development of 147 homes is nearing completion on former marshland in the Thames estuary in Kent.
The Barratt Homes project is in the Thames Gateway development region, one of the government's four designated growth areas for housing in south east England. The Gateway extends 64km eastward from London Docklands to Sheerness in Kent and Southend in Essex.
When developing marshland detailed and accurate site investigation is vital, but so is the data analysis, subsequent design of engineering works and ongoing monitoring.
Stats was recently appointed geotechnical consultant for the Barratt Homes development in Dering Way near Gravesend, which lies on the southern margin of the Thames fl oodplain. Brand Leonard Consulting Engineers is designing the estate roads and the houses.
The local geological map implies an increasing thickness of alluvium across the site from south to north, with chalk beneath. The chalk outcrops just south of the site.
It was formerly marshland, with drainage ditches running across it.
Much of the ground has been raised by about 2m to bring it safely above predicted flood levels.
Barratt is building an embankment to the north of the site, between 3.85m and 1.5m high, that will carry a distributor road to serve this and any future developments.
Stats was asked to carry out a full geotechnical review of the site, including a detailed investigation.
It did a feasibility study of types of foundation for the roads and houses and was also asked to design the distributor road embankment and foundations, including band drains, surcharging, basal reinforcement and specifi cation of fill material and its compaction.
Its brief included prediction and monitoring of primary and secondary settlement of soft alluvial soils beneath the distributor road embankment and housing areas, and monitoring pore water pressures in the soft alluvium.
Site investigation included a limited number of shell and auger boreholes, trial pits and window sampler holes primarily for obtaining samples for testing.
A comprehensive cone penetrometer testing (CPT) survey was carried out to delineate the extent and thickness of the soft alluvial deposits and the condition of the underlying chalk. Results indicated the soft alluvium varied between 2m deep to the south and 10m deep to the north east of the site.
Laboratory triaxial, oedometer and Rowe cell testing determined a range of soil settlement parameters for the alluvial deposits including coeffi ients of volume compressibility, consolidation and secondary settlement.
Given the highly organic nature of the alluvium, the coefficient of secondary settlement was particularly important as it allowed future settlements to be estimated.
This coeffi cient is less commonly used than the other settlement coeffi cients but is simply the slope of the secondary settlement curve (normally straight) per unit thickness of sample on a log time plot.
With the site investigation complete, a feasibility study of potential embankment foundations was carried out. The options were:
l A piled solution with geotextile reinforcement in the base of the embankment and piles founded in the chalk.
l A vibro concrete column (VCC) solution (there is insufficient confining stress for stone columns) with geotextile reinforcement in the base of the embankment and the VCCs founded on the surface of the chalk.
l Lightweight fi l material, such as expanded polystyrene, with excavation beneath the footprint of the embankment to balance vertical loads and limit long term settlement.
l Installation of band drains through the soft alluvial deposits down to the permeable chalk and subsequent surcharging to accelerate consolidation and possibe secondary settlement.
After a review of the options in September 2003, band drain and surcharging was chosen. This was because the method provided the most economic solution, while allowing the surcharged embankment to act as a haulage route for building materials and plant.
Design recommendations were fi nalised in February 2004. This provided specific details regarding the spacing of band drains and depth of surcharge required to achieve 95% primary consolidation within the desired time scale of eight months.
Various options were presented as consolidation rate can be varied by adjusting either the thickness of surcharge or band drain spacing.
The latter is particularly important, as the rate of consolidation is proportional to the square of the drainage path length. Clearly band drains can significantly reduce the drainage path length and thus markedly accelerate settlement.
For analysis a spreadsheet was produced which combined Terzaghi's theory of one dimensional consolidation with Barron's theory of radial consolidation.
In general, band drain spacing of 1m and 1.5m was adopted beneath roads and housing areas respectively, in each case a triangular grid pattern being chosen.
A surcharge thickness of 2m above fi nal ground level was used to accelerate settlements. A total of 29,692 band drains have been installed, an operation which took about 14 weeks using two rigs at times.
Based on the above spacing and surcharge, settlements across the site of between 100mm and 400mm after eight months of surcharging were predicted, depending on the thickness of alluvial deposits.
Along the distributor road, settlements were estimated to be between 200mm and 400mm.
A detailed assessment of estimated secondary settlement over 25 years was also made, based on secondary settlement coeffi cients from laboratory testing. This predicted secondary settlements of between 10mm and 40mm over the next 25 years after removal of surcharge.
The lower bound value is based on work by Mesri et al (1997) which indicates that surcharging of peat can limit future secondary settlement. The upper bound value assumes that future secondary settlement is not infl uenced by surcharging.
Details relating to the design of the distributor road embankment were also provided. This included specification of suitable fi ll materials (6F1) and their compaction, safe shoulder angles and basal reinforcement (Basetex 200/50 geotextile) required to provide embankment stability in the short term while the underlying soft alluvial deposits consolidated.
Further basal reinforcement (SS40 geogrid) was used in construction of a working platform/drainage blanket from which the band drains could be installed.
To monitor the ongoing performance of the band drains and surcharge, settlement plates were founded close to the base of the fill and vibrating wire piezometers installed across the site, predominately along the route of the distributor road.
Instrumentation was monitored fortnightly using a precise level and invar staff, providing data with an accuracy of between 0.1mm and 1mm. Vibrating wire piezometers measured the transient increase in pore water pressure after the application of surcharge and its subsequent reduction as consolidation of the soft alluvial deposits occurred.
In general, the measured settlements are much as expected. Settlements along the distributor road one year after surcharge was placed range between 120mm and 300mm.
Towards the west (SP01 and SP02 on graph) of the road where less organic material was present it can be seen that little secondary settlement is occurring.
To the east, however, where there is a considerable thickness of organic material, signifi cant secondary settlement is occurring. The secondary settlement coeffi cients backanalysed from these measurements are within 20% of that estimated by laboratory testing.
The vibrating wire piezometers reacted much as expected but provided little additional information to that provided by the level survey.
Pending satisfactory monitoring data, the surcharge in the distributor road area was due to be removed in early June so road construction could begin.
Shon Williams is director of ground engineering and Stephen Durham is senior geotechnical engineer at Stats.