1066
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
The response of the soil is not symmetric with respect to the
resonant frequency, which complicates the interpretation of the
data in terms of material damping. The logarithmic decay
curves were used to the scope. The results confirm once more
that suction affects the resonant frequency although to a lesser
extent than confining stress. As expected, equivalent viscous
damping decreases at increasing suction.
3.3
Theoretical and numerical modelling
While recent experimental efforts are dedicated to dynamic and
weathering behaviour, still modelling efforts are mostly
concentrated on developing constitutive laws for unsaturated
soil behaviour under static loads, especially in view of
engineering applications. Still a gap appears separating
advanced modelling frameworks from simpler laws, which are
suggested for straightforward application in the practice.
Vázquez et al.
(Spain) present a simplified model to predict
collapse upon wetting, based on oedometer test results on a
sandy-silty clay. They summarise their experimental results in
terms of an Instability Index, depending on applied vertical
stress, giving the amount of expected collapse as a function
initial suction and suction change. Although simplified, the
models depends on two parameters only, and can thus be
suggested for preliminary evaluation of collapse strain when a
one-dimensional geometric scheme can be applied in the field.
Zhao et al.
(China) discuss an interesting theoretical aspect
of unsaturated soil modelling approaches. Starting from the
observation that the critical state concept has been acting as a
cornerstone in the development of models for saturated soils,
they investigate the constraints leading to a thermodynamic
consistent definition of critical state for unsaturated conditions.
The Authors point out that for unsaturated soils variables
including the hydraulic state concur to a proper definition of
critical state. A thermodynamically consistent steady state is
reached when all the relevant static and kinematic variables,
including fluid pressures and volume fractions, reach their
asymptotic values, while only further deviatoric strain is
observed. Interestingly, they remind that critical state may not
be unique, depending on soil fabric, as already observed for
saturated soils with a dominant initial structure.
Advanced models, accounting for fabric, weathering,
degradation usually require a numerical implementation for
their evaluation, even before they are adopted to analyse
boundary value problems.
Fathalikhani & Gatmiri
(Iran)
present a coupled thermo-hydro-mechanical numerical
formulation of a model based on damage theory, developed to
analyse the effects of excavation in host geological barriers. In
the model damage is treated as a tensorial variable, accounting
for the directional crack pattern, while the effects of suction and
temperature are assumed to be isotropic. The model, developed
on both thermodynamic and micromechanical concepts, is
evaluated on a set of literature experimental data on a small
scale model of unsaturated bentonite, subjected to a heating and
a following relaxation phase. The consequences of suction and
temperature changes on damage are investigated during heating
and in the relaxation phase. The work highlights once more that
interpreting the multiphysics behaviour of soils is far from
being straightforward, and that even laboratory tests should be
careful analysed as boundary value problems at a small scale.
Following this line,
Kawai et al
. (Japan) present a Finite
Element model of compaction, using an elastic-plastic model
for unsaturated soils. A typical compaction stress history is
imposed, including the loading and the unloading stages. Void
ratio as well as suction are tracked along the stress path. The
simulation highlight that the state which is generally defined as
“as-compacted”, is the result not only of the loading stage, but
also of the following unloading path.
3.4
Modelling structures and infrastructures
The dependence of the as compacted state on compaction
history complicates the application of models for compacted
soils to full scale problems, as the paper by
Droniuc
(France)
suggests. In the latter contribution, a Finite Element analysis of
a model embankment made of fine grained soil is presented, and
the numerical results are compared to experimental
measurements from sensors installed in the model embankment.
The Author points out that, besides the choice of a proper hydro
mechanical model for the compacted soil, the analysis of an
embankment requires a careful investigation of the initial and
the boundary conditions. Swelling and shrinkage strains
promoted by soil-atmosphere interaction affect the state of the
material after compaction, and the heterogeneous profile of
suction and water content result in a heterogeneous response of
the system to hydro mechanical loads.
Sakai & Nakano
(Japan) present a preliminary study on the
effects of compaction on the dynamic performance of
embankments, motivated by design approach moving towards a
performance based concept in high seismic risk countries.
Samples of sandy materials, having different grain size
distributions, were compacted to different relative compaction
degrees, and subjected to constant water content test in a triaxial
apparatus. The results show that the density achieved during
compaction affects the liquefaction potential of the soil. The
experimental data are used to calibrate the model used to
perform preliminary analyses of the response of an ideal
embankment under seismic action.
Heyerdahl et al
. (Norway) introduce another complicating
feature in the assessment of infrastructures. Typically, a
consistent part of railway embankments are now about 100
years old. In spite of strengthening part of them with modern
criteria in recent years, still many of them are still working
under conditions which follow their original design. Prolonged
rainfall threaten the serviceability of the infrastructure system,
by inducing widespread damage of the embankments. In their
contribution the Authors provide an overview of damage
mechanisms, trying to classify the possible sources of
instability. Inadequate performance of culverts, water loads
from flooding and slope instability promoted by rainfall
infiltration are typically recognised as the main threatening
processes. The analysis of one case history, presented in the
second part of the paper, confirms that assessing an existing
structure, without detailed knowledge of its hydro mechanical
state and history, may be problematic, and that exhaustive
indication for reliable assessment may not be obtained.
As a whole, it seems that most issues related to the
assessment of existing structures and infrastructures concern
possible climate changes. In the contribution by
Makki et al
.
(France) the effect of differential settlements on masonry
building due to shrinkage promoted by extreme drought is
investigated, both on a prototype model and numerically
(Fig. 7). In the experimental test, the prototype was supported
by jacks, which were selectively removed to simulate
differential settlements. A 3D refined model of the masonry
construction reproduced well the observed displacements, and
allowed evaluating the structural behaviour of the structure
under possible action of drought.
Figure 7. Experimental masonry prototype and 3D numerical model
(from
Makki et al.
)
