Modeling Landslides with a Material Instability Criterion - Multiscale Geomechanics: From Soil to Engineering Projects - page 230

234 Multiscale Geomechanics

Moreover, it is assumed that the air pressure remains constant and equal to
atmospheric pressure and that χ parameter is equal to the degree of saturation. Hence,
we use the following relation:

σ ij = σ ij − S r u w δ ij [8.30]

with σ ij as the stress linked to strains by the elastoplastic constitutive model.

8.3.3. Landslide computation

For computational purposes, the elastoplastic model was calibrated on the
experimental tests performed by other people working on the LessLoss project.
Figure 8.8 presents calibration curves.

1400 0

-0.1

1200

-0.2

1000

-0.3

800

-0.4

600

-0.5

Po=300 kPa
Po=500kPa
Po=700kPa

400

-0.6

200

-0.7

0 -0.8

0 5 10 15 20 0 5 10 15 20

ε 1 (%) ε 1 (%)

Figure 8.8. Calibration of Plasol model for the Petacciato soil

It should be noted that the model employed does not permit us to describe softening.
Hence, only initial tangent mechanical properties and peak values are reproduced on
average. Table 8.2 gives the input values of the model.

As for the water retention curve calibration, no experimental results from the project
were provided. We, therefore, used experimental data from [FLE 02] soils having a
gradation more or less similar to that of the site. Figure 8.9 shows these results and
Table 8.3 provides the hydraulic parameters of the model.

Figure 8.10 displays the adopted mesh as well as the kinematic boundary conditions
considered. Concerning the imposed loading, our aim was to reproduce the influence of
the water table level on slope stability. For this reason, water pressure was imposed on
the left and right boundaries of the mesh. A hydrostatic pressure distribution, as defined
in Figure 8.11, was assumed. Below the point u w =0 ,the soil is fully saturated with
positive water pressures and above the soil is partially saturated with negative water