The influence of state on the capacity of driven piles in sands
The prime objective of the thesis was to study the influence of state on the capacity of driven piles in sand. The work was prompted by findings of recent laboratory based research on the mechanics of granular soils (Coop & Lee, 1993 and Jovicic & Coop, 1997) which showed that, as for clays, a principal controlling factor of soil behaviour is the stress-volume state. The work started in November 1996 and was funded buy the EPSRC under contract GR/L16590. Centrifuge testing was used to investigate the behaviour of driven piles in sands. A new actuator and strongbox package were designed and constructed taking full advantage of the centrifuge facilities. A fully instrumented model pile was developed, which consisted of eleven segments and allowed five independent measurements of shaft friction, four of radial stress and one of end bearing to be taken during installation. The pile was jacked into samples of two sands of varying densities at accelerations of between 50-200g, simulating piles of up to 70 min length and 3.2 m in diameter. The two sands used in the investigation were a carbonate sand and a quartz sand, chosen for the diversity of their geological origins and behaviour. A series of triaxial tests was conducted in order to characterise the behaviour of each sand, and in particular to locate precisely their critical state lines in stress-volume space. The centrifuge model tests showed that neither the end bearing nor the shaft friction could be adequately predicted using methods based on relative density of the sand such as APIRP2A (1993), but that the controlling factor for both was the state of the soil relative to the location of the critical state line, which should be quantified not only by the density of the soil, but also the effective stress level. The radial stress was found to be highest approximately seven pile diameters above the pile tip and not at the pile tip as assumed in recent design methods (e. g. Randolph et al., 1994 and Jardine & Chow, 1996). The friction angles mobilised on the model pile were found to be significantly lower than those that were measured by means of interface shear box tests, and it was found that the correct friction angle could only be measured by interface ring shear tests taken to very large displacements. An examination of the literature showed clearly that available field data support the new framework, although the data were both very scattered and very limited in extent, particularly because even the most extensive field tests rarely included even the basic laboratory tests required to apply the new method of analysis. The work therefore highlights severe limitations with many current pile design methods for piles in sands and suggests how new methods should be based on correct quantification of state.