Dewatering Case Study – Sirhowy Enterprise Way
1). Introduction
The scope of work was to design, install, maintain and remove on completion of dewatering, equipment, capable of dewatering and maintaining a dry stable excavation for the construction of a tower pad foundation of the Pont Dewi Sant Bridge, Blackwood, Wales. The suspension bridge has a span of 227m and its maximum elevation is 37m above the valley floor.
A short pump test was undertaken that enabled a final dewatering design to be made. A 14no deep well dewatering system was successfully installed and maintained required drawdown over an 8 week construction period.
2). Information
The tower was founded at 143m AOD with working ground level at ~ 148m AOD. The footprint of the pad foundation in 1:1 open-cut excavation ~ 50m x 22m.
Soil Information: BH 38, BH 38A, BH38B & BH 38C
Drawing No E4037-01 showing proposed erection layout.
3). Determining Aquifer Characteristics
The BGS solid geology map of the area (Newport, sheet 249) indicates mainly carboniferous sandstones form the Upper Pennant measure, with coal seams running through the area. The location is a glacial valley with flood plain of River Sirhowy. Study of the borehole information shows a typical profile:
| 0.0m – 0.3m | Top Soil |
| 0.3m – 2.5m | Superficial Deposits – clayey sandy GRAVELS |
| 2.5m – 5.0m | Glacial Deposits – slightly clayey sandy GRAVELS with cobbles |
| 5.0m – 9.0m | Glacial Deposits – very sandy sandstone, quartz and coal GRAVEL |
| 9.0m + | Bedrock – weathered Sandstone (Coal Measures) |
Artesian head within lower Glacial Deposits (5.0m to 9.0m)
Due to the artesian head in conjunction with a risk of excessive recharge form adjacent river and/or old mine workings. It was agreed an indicative design would be implemented and a short pumping test undertaken to finalise requirements.
Sirhowy Pumping Test
Pre test data indicated artesian level of 146.5m AOD. Anecdotal evidence suggests water level might be as high as 5m+ above GL. This is possible given the topography of the site. However standpipes were just artesian prior to pumping tests and may previously relieved some excess pressure.
The pumping well (BH 15) was constructed by Alan Metcalfe installing 250mmØ casing by cable percussive drilling methods. The boring encounted 3.4m of fine to course Gravels in two bands between 1.2-2.2m and 4.0-6.4m and terminated ~ 0.25m into bedrock at 8.7m. The screen was nominal 175mmØ uPVC 1mm slot-size c/w geo-sock wrap. The well was tested in February 2004 in accordance with BS 6316 (1982) by a series of step test, recovery and constant rate test. Data spread sheets are appended (Appendix 3). Water levels were recorded manually and with data-loggers. 4no. standpipes with installed at 12m to 17.5m radial distance (BH11, 12, 13, & 14) from the pumping well. These encounted varying conditions ranging from dominantly gravel to dominantly Boulder Clay. All bores terminated at bedrock (~9.0m).
The step test showed varying well losses as indicated by the well water level. There was some indication of well development during the initial stages of the test. The highest flow was 2.61 lts/sec; the well was pumped to below intake (~8m) in 26 minutes. The pumping test data is summarised on a drawdown versus log distance plot, attached. This showed good correlation between the piezometers and the well.
Analysis of a best fit line gives a transmissivity of 0.00048 m2/sec. If an average aquifer thickness of 4.0m is assumed it gives a permeability of 1.2 x 10-4 m/sec. No excessive recharge boundary was encounted, although it can be questioned if the test ever reached steady-state conditions.
4. Design of Dewatering System
With generated site data a dewatering design can be made, starting with a conceptual model that includes the following information:
Properties of the ground for Conceptual Model
| Aquifer type & properties - | Unconfined aquifer with artesian head to GL. Main aquifer are Glacial Gravel layers, however underlying sandstone has sub-artesian head to GL |
| Aquifer depth & thickness - | Depth varies in each borehole, occurring 1.0m to 11.0m BGL and variations in thickness from about 1m to 10m. |
| Presence of Aquitards & Aquicludes | Between the gravel layers occur laminated Clay & Silts which are likely to act as aquitards. A ~1.0m thick layer of mudstone often separates the lower Gravel aquifer from the bedrock, acting as a low permeable barrier and seepage face. |
| Radius of influence - | The River Sirhowy runs approximately 20m to the East of the site. A steep hillside runs down to the west and is believed to be source of artesian waters. Springs noted in hill side. |
| Groundwater Level - | Site area on flat flood plain. Groundwater level at ground level |
Other factors for Conceptual Model
| Geometry of Works - | Approximate 50m x22m area. GL at 147m to 149m AOD. Excavation to 143m AOD. |
| Dewatering Technique - | ~ 6.0m drawdown on limit of single-stage vacuum system. Wells will need to be drilled. Deep wells. |
| Time - | ~ 1 week to achieve drawdown and maintain for 8 weeks |
| Depth of Wells - | Gravels to be dewatered, wells to be toed into bedrock. |
| Environmental - | Discharge into river. Settlement Tanks, wells designed with appropriate filter pack to allow no removal of fines. |
Calculating Discharge (Q) on Drawdown (radial flow analysis)
To estimate steady-sate flow using established analytical methods Mansur & Kaufman (1962).
Full penetration of equivalent single well of unconfined aquifer fed by circular source
Q = 13.6 lts/sec (steady state) – F.O.S & allowance for storage 1:25 = 17.0 lts/sec
Conceptual Design: 10no x 10m deep wells (250mm¬a bore) equi-distance around perimeter of open-cut excavation.
Modelling of Dewatering System using cumulative drawdown analysis
Excel software developed by Project Dewatering Limited and the University of Newcastle was used for analysis of the interaction of different wells at variable locations to determine the pumping rate, position and number of wells required to achieve drawdown at a specific point.
Please find attached spreadsheet (Appendix 4) and plan of well locations.
The inputs into the spreadsheet
- Transmissivity (from pump test data) = 171 m2/day
- Storage coefficient (termed the Specific Yield for an unconfined aquifer) = 0.2
- Time = 3.5 days
Final Design: 12no x 10m deep wells (250mm¬ bore) equi-distance around perimeter of open-cut excavation.