Page 167 - Numerical Analysis and Modelling in Geomechanics

148 C.L.RAMSHAW AND A.R.SELBY
The pile-soil interface is modelled using a surface-based contact simulation,
with slip controlled by a Coulomb friction model with µ=0.1 (Mabsout and
Tassoulas, 1994). In order to simulate the horizontal stresses on the pile from the
soil, the false pile is expanded laterally into the soil by a predetermined distance.
(This avoids the problem of rigid body motion.) Once the horizontal stresses
have equilibrated, the vertical displacements computed from the pile model are
applied to the pile shaft nodes of the false pile and the soil nodes immediately
under the pile toe in a dynamic analysis with time steps of 0.001 s.
Outward transmission of ground waves is then computed using elastic soil
properties. Massarsch (1992) demonstrated that most of the energy is transmitted
in elastic waves beyond about a pile radius from the pile. Parametric studies of
site records of arrival times of waves at various distances from the pile indicate
that the dynamic soil stiffness exceeds the static stiffness. This has been
observed by several workers, e.g. Matthews et al. (1996). An appropriate FE
mesh would typically be a 50×50 mesh of 8-noded quadrilateral elements, each of
1 m×1 m. One of the more effective boundary strategies is also required, either
of grouped IEs, or expansion of the FE mesh to allow study of outgoing waves
without interference from spurious reflections.

Calibration of the impact driving model
Two sets of site data are next considered. The three-stage procedure described
above was used, and the parameters were adjusted in turn, so as to obtain the
best match between computed and site ground surface vibrations in both the
vertical and radial directions. However, it was anticipated that a perfect match
would never be achieved because of the assumptions of a pure axial impact,
shaft guide location, energy transfer and uniform elastic soils.

Impact driving at the M66
One of the rare occasions when site data included both pile head force-time, and
also adjacent ground surface vibrations was observed during dynamic testing of
driven cast in situ piles for bridge foundations on the M66 motorway near
Manchester.
Ten piles, 750 mm diameter and 21 m long, were tested by the SIMBAT
system (Stain, 1992). A concrete pile cap was cast and strain gauges,
accelerometers and an electronic theodolite target were attached. Signals were
recorded during blows from a 2.2 tonne hammer falling through 1.2 m. Ground
vibrations were measured by geophones placed on the ground surface at
distances of 5.5 m, 10 m and 16.5 m. Ground conditions comprised firm to stiff clay
soils from the surface to 10.5 m depth, underlain by dense sands. Static elastic
6
6
moduli were 13×10 Pa and 24× 10 Pa respectively, but a dynamic soil stiffness
6
of 150×10 Pa was found to be appropriate for dynamic small strains.