Page 322 - Numerical Analysis and Modelling in Geomechanics

F.BASILE 303

Figure 10.8 Group of 3 piles considered in comparison of methods.

Selection of soil parameters
In predicting the behaviour of pile foundations, the designer is faced with a
number of decisions, including the selection of the method of analysis and the
soil parameters to be adopted. It is crucial to recognise that the latter aspect is
generally of greater importance than the method of analysis, provided that a
soundly based method is employed.
Attention will be focused here on the estimation of the soil Young’s modulus
(E ), which is the key geotechnical parameter for pile deformation predictions.
s
The most reliable means of determining E is by backfiguring from the results of
s
full-scale pile load tests, using the same theory that will be used for the actual
deformation prediction. However, this is not always possible, at least in the
preliminary stages of design, and hence resort is made to the results of laboratory
or in situ soil tests.
The PGROUPN analysis is based on a non-linear hyperbolic interface model.
For this kind of analysis, previous experience has shown that the initial (“low
strain”) value of E may be successfully employed in the prediction of the initial
s
stiffness of the load-settlement curve of pile foundations (Poulos, 1989;
Randolph, 1994; Mandolini and Viggiani, 1997). The use of an initial tangent
soil modulus represents an advantage over a purely linear analysis which
requires a secant value of soil modulus, relevant for the applied load level.
Indeed, selection of an appropriate secant modulus is by no means
straightforward, whereas the initial modulus is a more reproducible quantity.
Some indication of the typical ratio of secant modulus to initial modulus as a
function of the applied load level has been presented by Poulos et al. (2001).
It is important to recognise that the value of E for the soil in the vicinity of the
s
pile shaft will be influenced by both the loading of the pile and the installation