Abstract Eggborough and Ferrybridge coal-fired power stations dispose of some of their ash residue by pumping it as a slurry to Gale Common where it settles out in lagoons.
In order to provide additional protection against seepage into groundwater, secondary HDPE liners were installed in the lagoons on top of settled ash.The ash is generally formed of silt-sized particles and the foundation for the secondary liners was expected to be very soft under the wet conditions. The design and construction of the liners is described, including a review of construction in similar conditions and a site trial to determine safe methods of construction.
The Gale Common ash disposal scheme in North Yorkshire is the largest of its type in the UK and disposes of 49M. m 3of pulverised fuel ash (PFA) from Eggborough and Ferrybridge C coal-fired power stations, as well as about 16M. m 3of spoil from the nearby Kellingley colliery.About 1Mt of PFA is pumped to the site annually as a slurry from the power stations some 5km and 7km away. Some of the slurry is processed in a vacuum filtration plant to produce conditioned PFA for use in embankment construction, the remainder being discharged to lagoons where it is allowed to settle.A site plan is shown in Figure 1.
Stage one of the scheme started in 1967 and is now nearly complete. It covers an area of over 160ha and consists of an embankment (crest length 2.2km) containing two lagoons separated by a narrow bund. The lagoons were raised alternately in 3.3m lifts and the supernatant water discharged via outfall towers into a culvert below the embankment to a nearby river.
As stage one neared its final crest level, the volume of materials to be disposed of began to exceed that required for embankment construction. Thus, construction of stage two embankments was able to commence using the excess of conditioned PFA and colliery spoil well before lagooning had to start, in 1986.
Construction of stage two is controlled by the rate of material supply (as was stage one) and it was necessary to achieve two targets before stage two could become operational. First, the banks had to be completed to level 13.1m OD,6m above ground level, in order to provide sufficient head to drive supernatant water to the settling ponds which are located near the return water pump house, some 2,500m from the outfall towers. Secondly, to construct the banks to level 19.7m OD so that their decreasing width with increasing height and the rate of supply of bank building material could support construction at the same rate as the lagoons are filled.
The requirement to build to a level of 19.7m ahead of lagooning explains the inner profile of the stage two embankments which have to be freestanding up to this level, but adopt a fir tree configuration at higher levels. In order to provide stability to the steep internal slopes, as their height was increased, it was necessary to partially fill the lagoons with water, which occurred between 1988 and 1989. The overall scheme is described further by Haws et al and Dennis et al and the form of construction of the embankment is shown in Figure 2.
The site was a flat area of marshy ground and is underlain by several metres of lacustrine and alluvial clays, sands and gravels, over bedrock. The solid geology comprises Carboniferous Middle Coal Measures at depth, overlain by Upper and Lower Magnesian Limestone and Permian Marl and partly by the Sherwood Sandstone Group.Details of the geology of the site and geotechnical characteristics of the foundation soils are given by Taylor et al.
The Magnesian Limestone, Sherwood Sandstone Group and basal sand and gravel are important aquifers. For stage one, the superficial clay and silt layers were relied on for prevention of potential polluting seepage from the lagoon base. Observation wells around the perimeter of the lagoons were monitored and in 1984 no adverse effect on the groundwater was reported by Brown and Owens. In view of the inability to prove complete continuity of the clay and silt blanket under stage two, a protective low density polyethylene membrane over the whole lagoon base was installed in 1986, and forms the primary liner.
To expedite draining down of the stage two lagoons upon completion of slurry placing, a dendritic pattern of drains, which are controlled by valves at the outlet culvert, was laid on top of the membrane. Initial filling of the lagoons with raw water, mentioned above, also protected the membrane from degradation and misuse.
Small heaps of colliery pressed tailings, a fine-grained waste product, were used to hold down the membrane from wind action during construction. In spite of this, a considerable amount of air was trapped under the membrane and it was necessary to carefully vent the 'whales'which arose as the lagoon filled.
Waste disposal licence
To comply with licensing requirements a secondary liner was installed as shown in Figure 3.Since the lagoons could no longer be drawn down without affecting the safety of the internal slopes, it was necessary to devise a means of installing the secondary liners at a level above the lagoon base commensurate with safety requirements and agreed with the licensing authority.
In order to do this it was proposed to fill the lagoons with hydraulically placed PFA (slurry) and subsequently draw them down using the dendritic drains.The settled PFA would then support the internal slopes and a dry surface crust could be formed over the lagoon as the internal water level slowly drained.
However the drainage conditions are relatively poor.The existing primary liner would trap water to its top level of +10.5mOD and the dendritic drains, at some 56m centres, might have limited value in draining the 4m, or more, thickness of settled ash within the timescale envisaged.
Review of previous experience
There is only limited experience of construction over settled ash at Gale Common and to provide some confidence that the method was feasible, a search for case histories was carried out. In addition, trials on fully saturated undrained settled ash were undertaken in the existing lagoon D of the emergency ash disposal area to provide data as an upper bound case for construction methods and to allow tenderers to price the risk element associated with the secondary liner installation. The trials were undertaken during the design period and were based on recommendations of geogrid manufacturer Netlon following its experience at Pomona Dock , Manchester.
The search for data showed that the Longannet and Cockenzie ash lagoons are vegetated only and that there is no experience readily available within Scottish Power Ash Sales of construction over PFA lagoons.
Personal communication with Medway Ports Authority identified that Lapwell Bank at Sheerness had been hydraulically filled using PFA and that areas that remained ponded, and thus saturated, were extremely difficult to work over. Drainage measures with a sand fill surcharge were used to improve the ponded area. It is used now as a car park and the only structures on it are lightweight portable cabins.
Boas describes a coastal development scheme using hydraulically placed PFA. The fill was placed behind bunds to 1m above sea level. It was then covered by a drainage blanket and a further 4m of PFA placed by dry methods.
Loading and field vibration trials are reported by Humpheson et al for building foundations on PFA-filled clay pits at Peterborough. Extensive testing of the fill was carried out but construction methods are not described in detail. Groundwater levels were generally at least 1m below the surface.
Concern was expressed about the vibratory effects of construction traffic and in one test the groundwater was artificially raised locally around the foundation, to surface level.However the vibrating source appears to have been distant from the foundation on a 500mm Type 1 fill over the PFA. The groundwater level at the vibrating source and the local effect at this point are not described.
The experience at Gale Common relates to trials carried out in 1982 to construct two embankments of some 10m crest width, one in shale and one in conditioned PFA, about 60m into lagoon A of stage one, to test final capping construction methods. Both embankments were constructed by normal earth moving methods and no problems were experienced in either material.The groundwater level fell from 4m below the surface to 7m below over the trial period.
Further work was carried out in 1990 in an effort to extend the life of the stage one lagoons and to make economies in the final capping required to conform to the final profile of the planning consent.This was to be achieved by constructing bunds to retain further lagoons within the existing lagoons, in place of dry filling. Site investigation and testing was carried out and initial construction was started as described by Martin et al.However the method was later abandoned because of programme constraints and lack of remaining lagoon capacity to accept the PFA produced by the power stations.
The limited case histories described above and the experience gained at Gale Common confirmed that it would be feasible to install a secondary liner over the settled ash in the stage two lagoons, although great care would be necessary during construction.
The previous experience in 1990 at Gale Common had indicated that a 1m thickness of unreinforced shale would provide a satisfactory working platform on settled ash, which had been selectively deposited by canalising the discharges into the lagoon.
This canalisation method of operation meant that the coarser sandier fractions settled first in the area to be reclaimed.However, for the more rapid construction required in stage two the use of geogrid reinforcement was proposed.The trials proved the feasibility of constructing a satisfactory working platform and determined the most cost-effective, safe method of construction.
An access bund was commenced by pushing shale out into lagoon D from the delta of material deposited at the slurry inlet.A crater forms underneath the mouth of the inlet pipe in the dumped material. The slurry subsequently flows over the dump into the lagoon and thus the delta progresses.Above water level the profile of the deposited ash is shallow and below it is much steeper.
The delta forms an extensive beach and progresses towards the outfall towers, with the slurry discharging into the lagoon via channels in the delta.The process of delta formation is described by de Groot et al and Blight, among others.
As the construction trial progressed wheeled dumptrucks were driven on to the promontory to deposit shale ready for spreading. One dumptruck, upon starting to empty its load, sank into the shale up to its chassis as the underlying settled ash liquefied.
The trial was suspended and only recommenced when a number of conservative working practices were adopted.These included thickening the access bund to 1.5m deep, prohibiting the use of wheeled vehicles on the trial sections and test loading each panel to prove its bearing capacity before allowing plant on top of it. These practices would have been extremely restrictive for the work on stage two lagoons, but might have been needed if particularly difficult conditions were found in certain areas of the prototype work.
The details of the trials are shown in Figures 4 and 5. Two promontories were constructed in successive panels approximately 10m square.Each panel was built by placing shale using a long arm excavator working from a previously placed and proof loaded section.The shale was built up in layers approximately 300mm-500mm thick. Compaction was by light tamping with the excavator bucket and later in the programme by a remote controlled pedestrian roller.Geogrid reinforcement was installed manually.
Each panel was loaded statically to 3t/m 2using shale on plywood boards over a 3m square in the centre of the panel. Testing was carried out a minimum of 12 hours after completion of construction of each panel.
Simple measuring rods were installed in the panels and were used to monitor movements by precise levelling at the panel/ash interface during construction and after any changes in loading or other external conditions, eg change of water level in the lagoon. Settlement under the self-weight of the panels was generally 16mm-26mm occurring overnight in the rest period between end of construction and testing.Little additional settlement occurred if testing was delayed.Under the test load, further settlements of up to 8mm were recorded, with some rebound after removal of the test load. There were insufficient data to establish longer term trends or relationships between settlements and varying loads over short-term periods.
The trials concluded that it was possible to construct a shale blanket over fully saturated ash using both reinforced and unreinforced shale.Approaching the lagoon waterline, it was impossible to continue building the blanket without using geotextile and geogrid reinforcement.Shale must be placed on the geotextile and geogrid layer to avoid tearing the material.
Geotextile was required as a separating layer between settled ash and shale and geogrid could not be used alone at the interface without punching into the settled ash.
Loading the fully saturated ash caused it to liquefy and excess water had to be allowed to drain away. Subsequently the ash was denser at formation level.The stone drainage layer in test panel G1 (Figure 5) greatly assisted drainage but was expensive and provided no significant advantage in construction time compared to the other panels where drainage appeared to occur overnight.
The trial panels were not tested to failure because of the safety risk and thus it was not possible to determine the optimum means of forming the prototype blanket.Where the settled ash conditions were very soft, a geotextile separating layer reinforced by geogrid was valuable in allowing construction to proceed.
The design of the shale blanket was based on the method ofMilligan et al and design parameters are given in Ta b l e 1 .The design method postulates that the relation between the required shear stress factor, a , acting on the subgrade (settled PFA) and the bearing capacity factor, N c, for the subgrade is linear and can be expressed as a= (Ka- Kp)*. *D2/(2*su*B7)+Nc(Ka*Ln(B'/B)/ tan BETA - tan d ) where D = depth of shale (1.0m) B = half width of wheel track (150mm) BETA = angle of load spread through fill (2V:1H) B'= half width of distributed load at fill/subgrade interface (650mm) . = unit weight of fill (20kN/m 3)Ka, Kp= lateral earth pressure coefficients su= undrained shear strength of subgrade (8kPa) d= mobilised angle of friction between fill and wheel load (0degrees) This factor is plotted on the interaction diagram for the shear stress and bearing capacity of the subgrade as the required force, shown in Figure 6. The subgrade resistance envelope is g iven by a= 1 for 0 Nc 1 + p /2a= 0 for Nc= 2 + pNc= 1 + (p /2)cos-1a+ (1-a2)1/2 The required force intersects the resistance envelope at Nc= 4.92, a = 0.2 resulting in a resistance force Puof 170kPa and a reinforcement tension, T of 1 kN/m based on Pu= Nc*su(1+ Dtan BETA / B) T = a suB' This resistance is greater than the shale fill bearing capacity, P, of 58kPa and ratio Nc= PB/(s uB') = 1.67 which is also plotted on Figure 6 and shows that failure of the shale blanket was the critical design criterion.Only nominal geogrid reinforcement was required for the subgrade.
In accordance with the results of the site trials and the design, three types of shale blanket were defined, as shown on Figure 7, depending on the condition of the underlying settled ash following drainage. The condition of firm, weak and very weak subgrade were determined as Firm: drainage of settled ash had occured after draw-down of the lagoon, where settled ash had been deposited above general lagoon water level Wea k : some drainage of settled ash Ver y w e a k : settled ash fully saturated In order to protect the secondary liner from scratches and punctures during construction the shale blanket was specified to be blinded with a 150mm layer of sand. To provide a weighting layer and prevent air trapped under the liner from floating it up when the lagoon was flooded, it was covered with a 200mm layer of sand and a 750mm layer of shale. Sand was selected in preference to conditioned PFA because it would permit working in a saturated condition during wet periods which is not the case with PFA, a silt sized material.
Discussions during the licensing procedure indicated that a 1.5mm thick high density polyethylene (HDPE) membrane would be acceptable as a secondary liner. On this basis a state of the art specification with detailed quality assurance requirements was prepared in accordance with US practice adapted to suit BS5750.Design of the membrane at the edges of the lagoons and for bridging soft spots followed the recommendations of manufacturers and Koerner.
The placing of shale is restricted to six months during the summer, although conditioned PFA may be placed year round provided that it is not frozen or saturated.Hence it was planned that the secondary liner in each lagoon, covering an area greater than 70,000m 2, would be installed in consecutive years. This fitted in with the alternating sequence of lagoon filling. It also gave a reasonable installation period and permitted some delay if there was an unduly wet summer.
The initial filling of lagoons was expected to leave areas of incomplete filling, particularly around the outfall towers. Thus provision was made to infill these areas by dumping conditioned PFA as the lagoons were drained in preparation for construction of the shale blanket.
At the end of 1993, slurry was directed into lagoon E of stage two to commence filling to the agreed level.Six weeks of draining down was permitted and construction of the shale blanket began in April 1994. It was found that there was no requirement for using conditioned PFA for low areas.
In addition, the blanket was successfully installed over its whole area without using layers of geotextile and geogrid. There had been sufficient drainage to create a firm subgrade. Lagoon F filling began in mid-1995 and only four weeks of draining down was available prior to construction start of the shale blanket.
The only change made to the design was to accept the contractor's request to change the sand layers to conditioned PFA which had been successfully used on another site, albeit for a much smaller area of membrane liner.
The shale working platform underlying the HDPE membrane was constructed by end tipping shale away from the working face.The shale was spread and levelled by bulldozer using a cascading face technique.Compaction was achieved by trafficking.At the end of each shift the shale was rolled and dressed to allow rainwater run off.
Close to the outfall area the settled ash was softer and platform construction caused it to heave, like bow waves, just in front of the leading edge of the platform. In this area it became more difficult to assess the placed thickness of shale platform.On a few occasions items of plant sank through the platform and became bogged down. These and other suspected soft spots were treated by surcharging with shale mounds.Shale placement was continued using excavators at this stage.The use of geotextile and geogrid in a single layer was necessary to create the shale platform in the outfall area.
At the outfall tower the heave of the settled ash had to be reduced by excavation to allow the final placement, to the required level, of the 1m depth of shale platform.All such work was carried out from the as-built platform.
The PFA bedding layer was planned to be laid and rolled only a little in advance of lining operations in order to minimise the area of damage should it have rained.Advantage was taken of fine weather to increase the area of PFA laid ahead of the liner and on only one occasion was it necessary to repair the effects of rainfall run-off from the HDPE liner.One disadvantage was that the large area of exposed PFA gave rise to considerable dust nuisance in windy conditions.
There were no particular problems with the HDPE liner except that the very hot weather gave rise to considerable expansion, causing ripples in the membrane. The necessity for one piece of housekeeping became apparent in that staples used to attach the liner roll to its cardboard former should be disposed of safely and immediately after cutting them off the roll. Failure to do this initially caused small pinpricks to the liner which then had to be repaired.
The upper protection layer to the liner comprised 300mm PFA which was the minimum thickness the contractor would lay in order to guarantee the performance of the liner.This layer was subsequently covered with 650mm depth of shale.
The PFA was placed by excavator and was advanced about 5m ahead of the shale.The excavator was trafficked across the PFA to compact it.Fresh PFA was dumped on to compacted PFA and great care was taken to avoid contamination of it with shale. Only the rearmost axle of the pair on the trailer section delivering the PFA was allowed on to the compacted PFA with the other axle on the shale. Excavators were used in preference to bulldozers for laying the PFA in order to avoid folding the ripples.The operators were able to deal with all of the small ripples (less than 100mm high) and most of the large ripples (up to 200mm high). It was expected that the ripples would contract out of the sunlight.The remaining large ripples were walked out to the nearby perimeter.
The upper shale layer was end tipped close to the working face and spread and levelled by bulldozer. Compaction was by trafficking of plant across it.
The construction process was independently audited as required by the waste disposal licence and a final report was prepared certifying satisfactory completion of the lining system.
A method was proposed for installing a secondary HDPE geomembrane liner over lagoons of settled ash which had little time to drain. In order to provide a suitable platform for the installation of the liner, a shale platform was proposed, based on previous small scale experience. The design method proved conservative in practice, possibly because of a safe choice of design parameters and better than expected drainage of the underlying subgrade.Construction was carried out to a rigorous QA programme and satisfactory completion to specified requirements was certified by an independent audit team.
The author is grateful to National Power for permission to publish this paper and to N King and T Smith for production of the figures.
Baker LH and McPherson DH. Cockenzie and Longannet power stations: novel features in the design and construction, ICE proc, Aug 1970.
Blight GE. The master profile for hydraulic fill tailings beaches, proc ICE, Geot Eng, Jan 1994.
Boas A.The design and construction of a coastal development scheme, Ashtec '84,2nd intnl conf on ash technology and marketing, London, 16-21 Sept, 1984.
British Standards Institution. Quality systems, BSI, London, 1987, BS5750.
Brown J and Owens PM. General and environmental aspects of ash disposal allied to land reclamation, 2nd int conf. on ash technology and marketing, London, 16-21 Sept, 1984.
Dennis JA, Hillier DH and Moggridge HT. Gale Common ash disposal scheme - planning, environment, operation and restoration, proc 6th BDS conf on the embankment dam, 1990, Thomas Telford, 1991.
Giroude JP and Peggs ID in Bonaparte R (ed).ASCE, Geotechnical special publication no 26.
Groot MB, Heezen FT, Mastbergen DR and Stefess H. Slopes and densities of hydraulically placed sands in Van Zykl DJA & Vick SG, eds, ASCE, Geotechnical special publication no 21.
Haws ET, Martin PL, and Orange-Bromehead RAN.Gale Common ash disposal scheme - concept, design and construction, proc 6th BDS conf on the embankment dam, 1990, Thomas Telford, 1991.
Humpheson C, Simpson B and Charles JA. Investigation of hydraulically placed PFA as a foundation for buildings, in ground movement and structures, ed J Geddes, intnl conf, 8-10 July, 1991.
Koerner RM. Designing with geosynthetics, 2nd edn, Prentice Hall, 1990.
Martin PL, Vaciago G, Somers NR, and Burbidge HT. Additional lagooning on stage one of the Gale Common ash disposal scheme: investigation, design and preparatory works, Proc 1st environmental geotechnics conf, Vancouver, 1994.
Milligan GWE, Jewell RA, Houlsby GT and Burd HJ.A new approach to the design of unpaved roads, parts I and II, Ground Eng ineer ing, Apr & Nov, 1989.
Taylor RK, Barton R, Mitchell JE and Cobb AE. The engineering geology of Devensian Deposits underlying PFA lagoons at Gale Common, Yorkshire, QJEG, vol 9, 1976, 195-216.