There are numerous references stating that the capacity of a pile driven to refusal may be governed by the structural capacity of the pile (for example Tomlinson, 1994).There are, however, few published case histories in the literature.
This article describes the application of driven tubular piles on a site in Toxteth, Liverpool, where made ground overlies weak Sherwood Sandstone (formerly known as the Bunter Sandstone). Figure 1 shows that the site lies close to the fault-controlled boundary between the Mercia Mudstone and the Sherwood Sandstone.
A ground investigation showed typically a 0.5m to 3m thick layer of loose sand and brick rubble fill overlying the sandstone. The sandstone was described as fresh, weak to moderately weak, being thickly laminated with closely spaced horizontal discontinuities. Point load and uniaxial compression tests were carried out on core samples and the results of these tests are summarised in Figure 2. The point load test results shown on this figure are given as an equivalent uniaxial compressive strength, q qc= 22. Is(50) MN/m 2after Brook (1985) The results suggest a spread in uniaxial compressive strength of between 3MN/m 2.This range is consistent with that shown by Hobbs (1974) for 'poorly cemented Bunter Sandstone' A number of cone penetration tests were also carried out on site; these generally met refusal within 0.5m penetration into the sandstone at an end-bearing pressure of 65MN/m 2, the capacity of the rig.
The design required a pile with a minimum safe working load of 400kN. It was thought that tubular piles would provide an efficient, cost effective solution and would easily penetrate the brick rubble fill on the site.
A 140mm diameter tube with a 10mm wall thickness was selected. The pile design was based on the Engineering News pile driving formula, which experience had shown would form a reasonable basis for assessing the driving plant and the final set requirement in relatively free-draining materials such as the Sherwood Sandstone.
The formula suggested a driving energy of 10kNm and a final set of around 2mm/blow would produce the required working load with a factor of safety of 2.0.
Verification of the design was initially provided by dynamic load tests and finally by static test. It was expected that the majority of the load would be carried in end-bearing and that plugging of the pile would not necessarily have contributed to the capacity; at its ultimate capacity the compressive stress at the toe of the pile on its annulus was expected to approach 200MN/m 2.The piles were driven using a conventional 2t drop hammer. Typically, final set was reached within about 0.5m penetration into the sandstone. Three dynamic tests were carried out with a piling rig and 2t hammer at different locations.
These tests were interpreted using both CASE and CAPWAP methods and yielded ultimate pile capacities varying between 820kN and 961kN, of which between 13% and 32% was estimated by the analyses to be in shaft friction.
In reviewing the results, it was thought that the dynamic tests might not be representative because of the failure to mobilise permanent displacement at the toe of the piles much more than about 2mm.
Subsequently, it was decided to install a test pile that would be subject to both dynamic and static tests. In this case, a 5.2t hammer was used working on a 0.2m drop (Figure 3). A penetration of about 0.4m into rock was achieved before a set of 1mm per blow was reached; the overall length of pile from the ground surface was 2.5m.
Dynamic testing was carried out on the following day and analysed by both CASE and CAPWAP methods as before. In this case, however, the heavier hammer allowed the mobilisation of about 5mm of permanent displacement at the toe of the pile, and higher pile capacities were estimated than previously; the end bearing capacity was estimated as 740kN, and shaft friction as 620kN, giving a total capacity of 1360kN (Figure 4).
A 24-hour static pile test was carried out two days after driving the pile.The load was taken to 1000kN, or 2.5 times the working load on the pile. Load was applied by kentledge (Figure 5), and controlled using a simple hand-operated hydraulic jack and calibrated load cells. The loading sequence is shown in Figure 6, and the results in Figure 7.
As may be seen from Figure 6, the loads were held for six hours at 400kN in the first loading cycle and 600kN in the second load cycle. No creep movement was recorded in the final one and a half hours of both the 400kN and the 600kN load cycles. The peak load of 1000kN was held for 30 minutes; a creep rate of 0.08mm/hour was recorded over the last 15 minutes at peak load, suggesting that the pile had much capacity in reserve even at 2.5 times its nominal working load.
It is clear that the ultimate capacity was not reached during the static test, and that the static test produced a stiffer pile response than estimated during the dynamic test. It is thought that the difference between the pile stiffness during the two tests is due to rate effects on dilation and hence the mobilised strength; higher penetrations are required under dynamic loading to produce the same resistance as under static loading.
In these materials and for this pile type, therefore, the dynamic testing can be considered to provide a lower bound to pile stiffness. The load capacity mobilised during the dynamic test is related to the hammer energy, and movement achieved at the toe of the pile.
The dynamic load testing suggests that an ultimate bearing stress approaching 200MN/m 2, approximately 50 q 2, or 22 q c.The similarity between this and the point load test correlation from Brook (1985) is probably coincidental. It is, however, useful to make comparisons with other design tools.
The principal components of the capacity of open ended tubular piles driven into rock are thought to be bearing on the annulus of the tube, and the formation of a plug of destructured rock within the annulus. The plug may be either sliding within the annulus or the rock below it may be in a state of failure.
Pells and Turner (1980) provide the following lower bound solution for the ultimate base resistance, q u, of a closed end pile in a c-w material:
Assuming w = 45infinity, and q c= 4MN/m 2, the lower bound solution implies an end bearing resistance of 23MN/m 2.In this case, the predicted load is approximately one quarter of the estimated ultimate capacity of the pile.Vesic (1972) gives a cavity expansion solution which makes some allowance for the confinement provided when a pile penetrates into a c-w material:
qu= c F c+ q F qwhere F cand F qare a function of strength and a factor, I rr , based on shear modulus.
Where I rr = 150 to 200 (typical for the Sherwood Sandstone), w = 45infinity and F c= 40, the equation predicts q u= 80 MN/m 2.Although neither of the above formulae makes allowance for dilation, it seems that the Vesic (1972) approach might be useful in predicting the capacity of a driven tube that has plugged.
Baligh (1976) adapted the method in predicting the capacity of driven tubular piles in sands with curved failure envelopes.
As pointed out by Hight et al (1996), the performance of the plug is affected by pile diameter and dilation of the material shearing within the plug. It is suggested that in sands at least, stable plugs are likely to form where pile diameters are less that 700mm; it by no means certain that this could be extrapolated to weak Sherwood Sandstone.
It is clear, however, that driving small diameter thick walled tubular piles into the sandstone produces very stiff piles of high capacity, and that plug formation contributes to the ultimate capacity of the piles.
There are however some uncertainties surrounding the formation and stability of plugs, and care should be exercised in extrapolating this result to different sizes of pile and different formations. Our ability to predict the performance of piles driven to refusal in rock would benefit from further testing work and the publication of results of tests.
Baligh MN (1976). Cavity expansion in sands with curved envelopes. Jnl Geotech Eng Div, ASCE, vol 102, no GT11, 1131-1147.
Brook N (1985). The equivalent core diameter method of size and shape correction in point load testing. Int Jnl Rock Mech Min Sci & Geomech Abtr, vol 22, no 2, 61-70.
Hight DW, Lawrence DM, Farquhar GB, Milligan GWE, Gue SS and Potts DM (1996). Evidence for scale effects in the end bearing capacity of open-ended piles in sand. Proc 28th Offshore Tech Conf, Houston, OTC 7975, 181-192.
Hobbs NB (1974). Factors affecting the prediction of settlement of structures on rock with particular reference to the Chalk and the Trias. Settlement of Structures, Cambridge, Pentech Press, 579-654.
Pells PJN and Turner RM (1980). End bearing on rock with particular reference to sandstone. Intl Conf on Structural Foundations on Rock, Sydney, 181-190.
Tomlinson MJ (1994). Pile design and construction practice, 4th edition. E&F Spon, London.
Vesic AS (1972). Expansion of cavities in infinite soil mass. Jnl Soil Mech & Foundations Div, ASCE, vol 98, no SM 3, 265-290.