Unexpected behaviour of piles installed in soft clay for a research programme at Kinnegar near Belfast has led to the site effectively becoming the first geotechnical test bed in Northern Ireland.
Research at a Department of the Environment (Northern Ireland) site at Kinnegar, about 10km north of Belfast, has evolved over the last two years to almost inadvertently create the first geotechnical test bed site in Northern Ireland. The main parts of the continuing research programme are:
Lateral load tests and combined lateral and vertical load tests on instrumented driven precast concrete single piles
Site characterisation involving insitu and laboratory testing
Tension and compression loading on pile groups and single piles using 250mm square precast concrete piles instrumented to a varying extent. Pile groups consist of five piles.
The project team was awarded an ICE Research & Development Enabling Grant in January 1988 which provided a great boost to the research participants: Trinity College, Dublin; Lowry McKinney Piling; John Barnett & Associates and Imperial College, London. The DoE (NI) and Queens University, Belfast are also contributing to a site characterisation study with Trinity College.
Site geology was determined by boreholes and piezocone tests. It comprises about 2m of alluvial clayey sand and silt overlying about 7m of soft, high plasticity, sensitive clay known locally as 'Sleech'. This clay is underlain by a medium dense sand. CPT qc values in the upper alluvium are typically about 1.5MPa while those in the Sleech are in the range 150kPa to 250kPa and representative of a very soft to soft clay. The water table is 1m below ground surface.
Lateral pile tests
Two 350mm square precast concrete piles (L1 and AL1) were driven, 2m apart, to the sand layer at 9.5m depth. Both had centrally placed inclinometer tubes, each of which housed a string of electrolevels for monitoring pile displacement profiles. They also contained strategically located pressure cells in addition to electrical and vibrating strain gauges on the reinforcing bars.
The soil in front of each pile was removed to a depth of 0.85m to
reduce the effect of the stiffer near-surface material on the lateral
pile response. An axial load of 460kN was then applied to pile AL1
while leaving pile L1 with no vertical load. The 460kN (the total
weight of the loading frame and kentledge used) induced a pile
head settlement of 2.5mm and was maintained while subjecting
both AL1 and L1 to a maximum lateral load of 60kN, one day after applying the vertical load. The vertical load on AL1 was subsequently reduced to 250kN and the piles reloaded the next day to a maximum lateral load of 89kN, resulting in a lateral displacement at both pile heads of 45mm.
The mechanism used to load the head of pile AL1 (Figure 1) included rollers at the underside of the beam supporting the kentledge and a pin just above the pile collar. The system was designed to permit free rotation and lateral displacement at the pile head while allowing the vertical load to be carried centrally on the pile.
However, strain gauge measurements made above ground level indicated that friction had caused the application of a small (righting) moment to the pile head. It was fortunate that the instrumentation allowed this moment to be measured reliably, and allowed for it when interpreting the test results.
One of the most interesting features observed was the ability of a vertical load, by transfer of positive skin friction, to lead to a
stiffer response of the soil close to ground level. For example,
the pile head movement of pile L1 at a lateral load of 60kN was 24mm
and that of pile AL1, even after correcting for the righting
moment induced by the pile head mechanism, was only 8mm. Such
a difference in movements could not be explained by natural variability of the soil.
While a more complete description and initial interpretation of these test results is provided in Lehane et al (1999), the following findings have immediate implications for the design of laterally loaded piles:
Lateral pile load tests are likely to indicate a lower lateral stiffness for the soil than would be developed under combined loading in service. However, as movements increase, the presence of a vertical load may contribute to a sudden collapse
The response of a pile to lateral load is critically dependent on
the characteristics of the soil within the top five pile diameters below the ground surface. Improving the soil stiffness in
this region could be considered as a cost-effective option at the design stage.
The assumption of either zero or infinite tensile strength for concrete when estimating the flexural rigidity of a laterally loaded reinforced concrete pile can distort the interpretation of lateral pile load test results.
Side friction (particularly for square precast piles) plays an important role in controlling the response of piles at conventional working displacement levels.
Pile group tension test
Phase 1 of the pile group experimental programme at Kinnegar involved tension tests on a group of five piles and a single reference pile. All piles had 250mm square precast concrete sections and were driven to 6m depth.
The group comprised four outer piles at a radius of 720mm from the central pile. The central pile was driven first and was subsequently retapped following driving of the outer piles. Pneumatic piezometers showed that maximum excess pore pressures generated by group installation (measured at depths between 2m and 5m) varied from 1.2'v0 at 1m from the group centre to 0.4'v0 at 2.5m from the group centre ('v0 is the free field vertical effective stress). Within two months of group installation 85% of all excess pore pressures had dissipated.
Group tension loading took place a year after installation using
the arrangement shown in Figure 2. A load cell was fitted to
each pile head and welded to the steel pile cap, with the cell also providing the anchorage point for Dywidag bars that were grouted
into the pile heads. The tension load was applied by pulling two high strength bars bolted to plates on the underside of the pile cap. The
two jacks were seated on a deep beam and the compressive supports
to the applied tension load were provided by slip-coated compression piles that were driven to the sand layer at 9.5m depth and linked
with a concrete cap.
The tests were load controlled, with creep rates falling to below 0.24mm/hour between load increments. The overall group response under tension loading and that of the single control pile is summarised in Figure 3, which shows that:
The group stiffness (kN load per pile, per mm uplift) is lower than that of the single reference pile. The centre pile, pile 3, displayed the stiffest individual response of the five piles in the group.
The ultimate pull-out capacity of group piles is typically some 15% less than that of the single reference pile. Pile 1 did, however, have an ultimate capacity similar to the single reference pile.
Although not evident from Figure 3, the centre pile attracted
50% more load than each of the outer piles, up to 75% of its ultimate capacity. At this point, the rate of load increase in the centre pile
reduced significantly and, at ultimate conditions, all piles (except
for pile 1) effectively carried the same load. The effects of rigidity of
the pile cap on the response of the group are currently being
Phase 2 of the testing in Kinnegar, which will be completed this summer, comprises a similar set of tests to those outlined above but loading will be in compression rather than tension. The 'compression piles' have been fitted with strain gauges and pressure cells.
The Kinnegar programme of field and laboratory research has indicated interesting and unexpected aspects of pile behaviour in soft clay. Further useful results should emerge. The research programme has involved a successful and constructive combination of research institutions and industrial partners.
The authors would like to acknowledge the significant contribution of the following research assistants: Bryan McCabe and Declan Phillips at Trinity College, Dublin, and Adam Pellew at Imperial College.
Lehane BM, Phillips DM, Paul TS and Horkan E (1999). Instrumented piles subjected to combined vertical and lateral loads. Proc. XII European Conference on Soil Mechanics and Geotechnical Engineering, The Netherlands, June 1999. Publication pending.