Numerical Modeling: An Efficient Tool for Analyzing the Behavior of Constructions - Multiscale Geomechanics: From Soil to Engineering Projects - page 299

304 Multiscale Geomechanics

is observed on the face of the retaining wall. With the Mohr-Coulomb criterion, this
area seems to bear the characteristics of an active zone. A more complete analysis of
other geometries, configurations, and meshes is available in the PhD dissertation of
K.C. Vossoughi [VOS 01].

9.5.2. Vertically loaded piles

The analysis of load transfers in vertically loaded piles necessitates a good
comprehension not only of the soil behavior, but also of the soil-pile interaction. In this
problem, the parameters such as the initial state before and after the installation of the
pile, the loading history and stress path, as well as the roughness of the interface, play
an important role. Moreover, it is important to analyze the cyclic behavior of the pile
and the interface, especially for studying the effect of installation.

In the framework of Sofia D'Aguiar's PhD thesis [DAG 08], several series of
numerical computations simulating non-displacement piles have been performed in
ordertosimulateandtoobtaintheresultsofphysicalmodeltestsperformedincentrifuge
by Fioravante [FIO 02]. Both rough and smooth piles have been tested in Toyoura sand
with a relative density of D r =0.93 . The properties of the interface are based on the
behavior of the sand and should take into account the roughness of the interface. As
these characteristics influence the friction mobilization at the interface, it is important
to develop a coherent strategy to identify the parameters of the interface model based
on the soil model and the roughness of the interface. Fioravante has performed shear
tests on sand-iron interfaces with two different normalized roughnesses ( R n =0.01
and 0.13 ; R n =

R t

d 50 , where R t is the depth of surface asperities). This test is similar
to a simple shear test where the normal stiffness is kept constant. Furthermore, the
Toyourasandbehaviorhasbeenstudiedbymanyresearchers[ISH 93,IWA 78,KOK 82,
KAB 91].

The comparison of experimental results with those simulated numerically is given
in Figure 9.25 for the Toyoura sand and in Figure 9.26 for the interfaces. Tables A.2
and A.3 in the annex in section 9.8 summarize the parameters used in the numerical
models. In order to have a global view of the multimechanism model's response and
the coherence of the chosen set of parameters for the Toyoura sand, interested readers
are invited to consult S. D'Aguiar's PhD thesis [DAG 08].

The roughness of the interface is taken into account by reducing the friction angle
φ pp from 30 to 16 degrees, by increasing the plastic compressibility β to reduce
dilatancy and by changing the form of the yield surface (parameter b ). Consequently,
for R n =0.01 and smooth interface, contractancy/dilatancy is very low in comparison
to the values for rough interfaces: R n =0.13 .