Geoscience Reference
In-Depth Information
276
Multiscale Geomechanics
sands and heavily overconsolidated clays. It is very difficult to know the position of
the critical state line and its distance compared to the line of virgin consolidation in
(
e − ln p
) space or the angle of friction at perfect plasticity. Similarly, the small size
of the elastic domain for granular materials makes it very difficult to determine their
elastic properties. On the other hand, despite all the experimental difficulties, if we do
manage to obtain consistent and reproducible results on samples made from remolded
materials prepared in the laboratory, it may still not be possible to obtain similar results
from unremolded soil sampled
in situ
. Therefore, opting for the strategy of identifying
model parameters, we used the results of many studies for their correlations between
intrinsic characteristics of the soil, such as the elasticity modulus, or those related
to the critical state and other characteristics often used in soil mechanics to classify
soils, such as Atterberg limits and the consistency index for clays and characteristics
related to the grain-size distribution for sands. Although these correlations help identify
parameters, a good knowledge of the physics contained in the model is absolutely
indispensable.
Given the non-linear character of the latter and the high number of parameters, we
are not sure to obtain a single set of parameters after calibration, especially if only one
type of stress path is considered. Using correlations and understanding the physical
model and its parameters is the only way to ensure proper calibration. To do this, we
must add that the mechanisms are directly related to the loading path experienced by
the soil. We first present the strategy of identifying parameters for remolded materials
for which many experimental results are available. Moreover, several authors have
published reference curves for different materials by conducting the synthesis of a
large number of experimental results. In practice, different types of tests are performed
in the laboratory to characterize soils. They concern essentially triaxial compression
or extension or shear tests under drained or undrained conditions. Resonant column
tests are performed for geotechnical earthquake engineering applications. Biarez and
colleagues propose reference curves for Hostun sand, depending on its density, and also
clays under monotonic loading. For cyclic behavior, the curves given by Vucetic and
Dobry [VUC 91] for clays and those of Seed and Idris [SEE 70] for sands are widely
used. In presenting the parameters' identification strategy, we try to be as general as
possible. Where the parameters must be identified by curve-fitting, the reference curves
are used. This strategy will then be applied to soils of well-studied sites to justify its
efficacy for the study of real problems.
9.4.1.
Classification and identification of the ECP model parameters
In this study, we focused on aspects of parameter identification for the simulation of
the behavior under cyclic loads. In fact, although this type of loading is routinely applied
in earthquake engineering, geo-structures regularly undergo loading and unloading. As
Search WWH ::
Custom Search