242 Pile groups under compressive loading
An important point to note in the application of soil mechanics methods to the design of
pile groups is that, whereas in the case of the single pile the installation method has a very
significant effect on the selection of design parameters for shaft friction and end bearing,
the installation procedure is of lesser importance when considering group behaviour. This is
because the zone of disturbance of the soil occurs only within a radius of a few pile diameters
around and beneath the individual pile, whereas the soil is significantly stressed to a depth
to or greater than the width of the group (Figure 5.1). The greater part of this zone is well
below the ground which has been disturbed by the pile installation.
Computer programs have been established to model pile–soil interaction behaviour from
which the settlement of pile groups and the loads on individual piles within the group can
be determined.
Some of the programs are
DEFPIG Non-linear continuum analysis using interaction factors
GAPFIX Non-linear continuum analysis, complete solution
M-PILE Simplified continuum analysis using interaction factors
PGROUP Complete linear continuum analysisIn the above programs soil behaviour is modelled on the basis of the theory of elasticity.
Poulos(5.1)states, ‘Despite the gross simplification which this model involves when applied
to real soil, it provides a useful basis for the prediction of pile behaviour provided that
appropriate elastic parameters are selected for the soil. A significant advantage of using
an elastic model for soil is that it provides a rational means of analysis of pile groups and
evaluation of immediate and final movement of a pile. In determining immediate move-
ments, the undrained elastic parameters of the soil are used in the theory, whereas for final
movements the drained parameters are used’. A useful comparison of the M-PILE and
PGROUP programs is given in the UK Department of Transport Publication BD 25/88.
The interaction factors depend on the geometry, stiffness and spacing of the piles and the
elastic modulus of the soil between them.
In view of the above reservations and the difficulties of obtaining representative values
of the undrained and drained deformation parameters (particularly the latter) from field
or laboratory testing of soils and rock, the authors believe that the equivalent raft method
is sufficiently reliable for most day-to-day settlement predictions. Nevertheless, it could
be convenient and time saving to use an available computer program particularly when
making studies to determine the effect of varying parameters such as pile diameter, length
and spacing.
In most practical problems piles are taken down to a stratum of relatively low com-
pressibility and the resulting total and differential settlements are quite small such that an
error of plus or minus 50% due to deficiencies in theory or unrepresentative deformation
parameters need not necessarily be detrimental to the structure carried by the pile group
(see also Section 11.1.4).
As an example of the relative accuracy of the methods Figure 5.4 shows a 44 pile group
where the piles spaced at 3 diameters centre to centre are taken down to a depth of 24 m into
a firm becoming stiff normally consolidated clay where the undrained shear strength and
compressibility vary linearly with depth. The group settlements calculated by the equivalent
raft method used the influence factors of Butler (Figure 5.19).