Recent advances in practical design of rock socketed piles in Victoria

Chris Haberfield and Ben Collingwood


Piles socketed into rock are a common foundation solution to carry large loads from buildings and bridges. They have experienced widespread use in Melbourne for many years and have been installed over a wide range of subsurface conditions. The load applied to the pile is primarily carried by the bottom part of the pile which is “socketed” into the rock. Rock sockets in Melbourne comprise siltstone/sandstone and Newer and Older Volcanics (basalt). Throughout the world, rock socketed piles have been and are still designed by many engineers by assessing estimates of allowable shaft and allowable base resistance, with little if any assessment of the displacement performance of the pile. Settlements are quoted based on experience rather than by calculation. The estimates of allowable shaft and base resistance are usually based on local ordinances, experience or empirical correlations with the uniaxial strength of the intact rock (e.g. see Figure 1).

The performance of most rock socketed piles at serviceability loads is dependent predominantly on the shear resistance developed at the interface between the concrete shaft and the surrounding rock, i.e. shaft resistance. The shaft resistance developed depends on many factors including the shaft diameter, the type, stiffness and strength of the rock and construction effects such as socket roughness, the thickness of smear zone or residual drilling fluid coating the socket walls and the pressure imposed onto the socket rock due to fluid concrete placement (e.g. Johnston, 1977; Williams and Pells, 1981; O’Neill and Hassan, 1994). Shaft resistance is very sensitive to a number of these parameters and as a result, socket performance can vary significantly from one site to another, even in the same rock type. This provides an explanation for the large scatter observed in published correlations between shaft resistance and rock strength; for example Williams, Johnston and Donald (1980) or Rowe and Armitage (1984). The use of such empirical correlations requires the designer to choose a design shaft resistance from within a relatively wide scatter and as a result a conservative assessment is usually made.