Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Pile disturbance in layered ground


Introduction by Keith Emmett, department of civil and structural engineering, The University of Sheffield.

Redevelopment of sites that have previously been used for domestic or industrial purposes is becoming increasingly necessary and desirable.

However, many of these sites contain contaminants in the upper soil strata. The danger of spreading these insitu pollutants is an issue facing engineers when designing deep foundations such as piles.

The Environment Agency has identified the risk of piling on brownfield sites and the creation of flow paths that it may cause, allowing contaminated groundwater to move through low permeability layers into underlying aquifers (Environment Agency, 2001).

There is a lack of research and monitoring data on contaminant migration through layered ground. Some preliminary work was carried out by Hayman et al (1993). Solid cylindrical and wooden piles were driven through a layered ground model containing a thick intermediate clay layer, equivalent to 60 pile diameters. Results showed that the clay would seal the edges of the piles, preventing the creation of preferential flow paths.

Boutwell et al (2000) also modelled driven piles in layered ground with clay layers equivalent to 12 pile diameters. Results showed sealing around the solid piles and an increase in vertical permeability for H-section and stone column piles. However the formation of the preferential flow paths remained unclear.

The aim of this paper is to understand how preferential flow paths are formed by driven piles in layered ground and establish a relationship between their impact on vertical permeability and the thickness of the clay layer.

Modelling strategy A 1/10th scale model simulating an overconsolidated clay stratum between two sand layers was prepared in the laboratory. Both the impact of the clay layer thickness and the pile type used on the formation of preferential flow paths were examined. The flow of water and the differential pressure across the model were measured before and after pile construction so that the changes in vertical permeability caused by each pile could be quantified.

Sample set-up Test cell and model pile For the purpose of this research, a test cell was designed to simulate field conditions and monitor the permeability changes in layered ground due to the construction of piles (Figure 1).

The soil specimen was self-contained inside a membrane attached to plates at either end. This ensured that water would only flow through the model and any variation between in flow and out flow indicated the existence of a leak. A constant head test was carried out for each piling scenario to establish the permeability of the specimen. For flow rates less than 50mm 3/s, constant flow tests were carried out using a GDS Instruments pump.

To evaluate the permeability of the model, inflow and outflow were constantly measured using volume change units. A transducer measured differential pressure across the length of the sample. Porewater pressures directly above and below the clay layer were measured with two transducers attached to 3mm diameter probes. These probes were manually pushed into the model after pile construction to monitor head losses caused by the preferential 'ow around the pile.

Three test cell lengths ranging from 350mm to 500mm were made to accommodate the varying soil thicknesses simulated. The cells included a 40mm diameter plughole through which the 25mm diameter piles were driven into the soil. The pile to test cell diameter ratio was 1/10, judged sufficient for the boundaries of the cell not to affect the deformations caused during pile construction (Huang, 1991; Hird and Moseley, 2000).

The cylindrical piles used in the experiment were made from stainless steel and the H-section piles from high strength commercial aluminium. The surface of the stainless steel piles was scoured with fine glass paper before every test to maintain a constant roughness, whereas a new H-section pile was made for each experiment.

Upper and lower stratum sands For the purpose of this research medium silica sand (1.18mm to 600mm, fraction B, Leighton Buzzard) and fine silica sand (600mm to 150mm, fraction C, Leighton Buzzard) were used to simulate the upper and lower strata respectively.

It was hoped that, after testing, the lower stratum could be sieved to quantify the amount of medium sand transported down by the pile. However, the segregation of particles during sample building, the crushing of particles during pile installation and the small quantities of material transported relative to the total amount sieved gave inconsistent results. These results are therefore not included in this paper.

Maximum and minimum densities (using BS1377: Part4: 1990:4.3 and BS1377: Part4: 1990:4.2), void ratios and the theoretical permeability of the two sands (Chapuis and Aubrtin, 2003) are listed in Table 1.

Aquitard layer Speswite kaolin clay was used to represent the intermediate aquitard layer.

The clay powder was mixed with deionised and de-aired water to form a slurry at twice the liquid limit (Sheeran and Krizek, 1973). The slurry was then placed and de-aired in two 250mm diameter Rowe cells where it underwent one dimensional consolidation. Cell pressures of 12.5kPa, 25kPa, 50kPa, 100kPa and 200kPa were used to consolidate the slurry, with transducers measuring the settlement at each pressure increment. Typical values of the moisture content, void ratio and theoretical permeability of the consolidated clay (Al Tabbaa and Wood, 1987) are presented in Table 2.

Test procedure Sample building The upper sand stratum was placed in three sub layers, each of which was thoroughly mixed with de-aired water to maximise saturation and compacted for 10 seconds using a Kango hammer.

Preconsolidated kaolin clay was then wire cut to the required thickness and placed carefully on top of the coarse sand layer, using a suction pad, to avoid sand grains being trapped around its edges.

The lower fine sand stratum was then loosely placed under water to avoid disturbing the rest of the sample and carefully stirred to remove trapped air.

Once the Rowe cell lid was attached to the test cell, confining pressures were applied to the specimen (effective vertical pressure of 30kPa and effective lateral pressure of 10kPa) before the cell was pivoted.

Effective vertical and lateral pressures were then raised and maintained at 100kPa and 80kPa respectively. A back pressure of 200kPa was gradually applied to dissolve any air trapped in the sample before pre-piling permeability testing was carried out.

Piling procedure Before piling took place, back pressures were reduced to zero and confining pressures maintained at test level (effective vertical pressure of 100kPa and effective lateral pressure of 80kPa).

The 40mm diameter plug was removed to allow the pile to be driven into the soil before being reconnected. Both cylindrical and H-section piles (25mm diameter and 25mm wide) were driven into the sample by a hydraulic ram at a constant rate of 5mm/s (Figure 2). Although the sand was under drained conditions, the drive rate ensured that the clay was under undrained conditions.

The confining pressures were then reduced by 50% so the pore pressure probes could be manually inserted into the sample without soil friction creating too much resistance (Figure 3). Finally, back pressures and confining pressures were gradually restored before post pile permeability tests were carried out.

Test results Figure 4 shows the net increase in unit gradient flow around the pile at full scale with respect to the ratio between the pile diameter and clay layer thickness. There is a distinct cut-off point at a clay layer equal to two pile diameters, where the preferential flow paths created by solid cylindrical piles are sealed by the aquitard. However, even at a clay depth equivalent to eight pile widths, H-section piles still create preferential flow paths through the impermeable stratum.

By looking at the cross section core images of each test, it is possible to establish the deformation mechanisms associated with each pile and identify the preferential flow paths that are being created.

Solid cylindrical piles The clay layer thicknesses in Figures 5a, 5b and 5c are equivalent to one, two and four pile diameters respectively.

Each figure shows the formation of a cone of upper stratum sand penetrating the clay layer, a key feature of how preferential flow paths are created by solid cylindrical piles. However, the cone is formed from a finite source of material, which is why clay depths greater than two pile diameters have effectively sealed the drainage path (Figures 5b and 5c).

A five stage theoretical model of how solid cylindrical piles deform the layered ground is shown in Figure 6:

Stage 1: There is an increase in effective stresses below the base of the pile as it is driven into the sand (White & Bolton, 2004).

Stage 2: These stresses increase to an extent where the sand overcomes the shear strength of the underlying clay and pushes through.

Stage 3: The initial penetration of the sand forms a cone and a plug of material trapped at the base of the pile.

Stage 4: As the pile is driven further into the aquitard, the sand at its base starts to mix with the clay due to the stresses in the zone below it being increased by the intruding object.

Stage 5: The mixed material is forced around the edges of the pile base and is deposited as a bulb on the lower clay layer boundary, where the stiffness of the underlying sand stratum is higher than that of the aquitard. Some of the upper stratum material trapped beneath the pile is carried through into the lower sand layer and, depending on its compressive strength, will eventually be crushed as the pile is driven deeper.

H-Section piles

The clay layer thicknesses in Figures 7a, 7b and 7c are equivalent to two, four and eight pile widths respectively. Each shows upper stratum sand transported through the clay layer between the web and flanges of the pile. Here, a more permeable zone is created allowing fluids to flow more rapidly through the aquitard.

Figure 8 illustrates five stages of how these preferential flow paths may theoretically be formed:

Stage 1: As the H-section pile is driven into the soil, coarse grained material is transported between its web and flanges.

Stage 2: The intrusion of the pile increases the effective stresses in the soil beneath it until they are sufficient for the upper stratum particles to penetrate the aquitard.

Stage 3: As the pile itself punctures the impermeable layer, the final shape of the sand cone is formed and clay starts to build up between the web and flanges.

Stage 4: Softer clays will have a tendency to coat the sand grains, whereas stiffer clays will provide a more resistant barrier to the amount of sand transported down by the pile. Some of the upper stratum material at the pile base will mix with the clay and form a bulb, which will be deposited on the lower aquitard boundary.

Stage 5: As the pile flnally penetrates the lower stratum sand, the abrasive nature of the particles will progressively remove all of the clay inside the web and flanges, minimising its sealing effect on the preferential flow path.

Conclusions Solid cylindrical piles cause large soil displacements. However, the sand cones formed in the clay layer are made from a limited source of material, allowing the preferential flow path to be effectively sealed at clay depths greater than or equal to two pile diameters.

H-section piles cause little soil disturbance around their outer edges.

However, upper stratum material transported through the clay between the web and flanges of the pile is sufficient to create preferential flow paths through an aquitard layer equivalent to eight pile widths.


Al-Tabbaa A and Wood DM (1987). Some measurement of permeability of kaolin, GÚotechnique 37, No. 4, 499-503.

Boutwell GP, Nataraj MS and McManis KL (2000). Deep foundations on brownfield sites. Prague conference, Prague.

Chapuis RP and Aubertain M (2003). On the use of the Kozeny-Carman equation to predict the hydraulic conductivity of soils. Can. Geotech. J. 40, 616-628.

Environment Agency (2001). Piling and penetrative ground improvement methods on land affected by contamination: guidance on pollution prevention. Bristol: Environment Agency.

Hayman JW, Adams RB and Adams RG (1993). Foundation piling as a potential conduit for DNAPL migration. Proceedings of the Air and Waste Management association Meeting, Denver.

Hird CC and Moseley VJ (2000). Model study of smear zones around vertical drains in layered soil.

GÚotechnique 50, No. 1, 89-97.

Huang AB (1991). Calibration chamber testing. Elsevier, New York.

Sheeran DE and Krizek RJ (1973). Preparation of homogeneous soil samples by slurry consolidation.

Journal of Materials JMLSA, Vol. 6, No 2, 356-373.

White DJ and Bolton MD (2004). Displacement and strain paths during plane-strain model pile installation in sand. GÚotechnique 54, No. 6, 375-397.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.