2. Modeling A#
2.1. Characteristics of modeling#
The modeling is \(\mathrm{3D}\) with a hydro-mechanical coupling in quasistatic non-linear mode.
In loading phase 1, the sample is brought to consolidation pressure \({\sigma }_{\mathrm{xx}}^{0}={\sigma }_{\mathrm{yy}}^{0}={\sigma }_{\mathrm{zz}}^{0}={\sigma }^{0}=-30\mathrm{kPa}\)
We use Hujeux’s law cyclic.
2.2. Characteristics of the mesh#
Number of knots: \(20\)
Number of meshes and type: 1 HEXA20 and 6 QUAD8
2.3. Tested sizes and results#
The solutions are calculated at point \(C\) and compared to the results obtained with GEFDYN. They are given in terms of isotropic pressure, plastic volume deformation \({\varepsilon }_{v}^{p}\) and mobilization factors, and summarized in the following tables:
\(3\mathrm{\cdot }P\text{'}\mathrm{=}{\sigma }_{\mathit{ij}}\mathrm{\cdot }{\delta }_{\mathit{ij}}\) \((\mathit{kPa})\)
(\(\mathit{kPa}\)) |
Code_Aster (\(\mathit{kPa}\)) |
GEFDYN (\(\mathit{kPa}\)) |
Relative error |
-15 |
-80.223 |
-80.194 |
0.038% |
15 |
-73.988 |
-74.078 |
-0.122% |
-15 |
-66.055 |
-66.250 |
-0.295% |
15 |
-52.625 |
-52.999 |
-0.707% |
-15 |
-45,539 |
-45.672 |
-0.292% |
\({\varepsilon }_{v}^{p}\mathrm{=}\mathit{trace}({\varepsilon }^{p})\)
(\(\mathit{kPa}\)) |
Code_Aster |
GEFDYN |
Relative error |
-15 |
-2.62e-5 |
-2.63e-5 |
-0.225% |
|
-2.80e-5 |
-2.78e-5 |
0.854% |
|
-4.48e-5 |
-4.43e-5 |
1.076% |
|
-7.11e-5 |
-7.05e-5 |
0.842% |
|
-1.13e-4 |
-1.11e-4 |
1.412% |
|
-1.85e-4 |
-1.85e-4 |
0.261% |
\(({r}_{\mathit{iso}}^{c}+{r}_{\mathit{ela}}^{s})\)
(\(\mathit{kPa}\)) |
Code_Aster |
GEFDYN |
Relative error |
|
0.00140 |
0.00138 |
1.329% |
|
0.00220 |
0.00217 |
1.406% |
|
0.00341 |
0.00337 |
1.246% |
|
0.00518 |
0.00510 |
1.663% |
|
0.00195 |
0.00189 |
2.947% |
\(({r}_{\mathit{dev}}^{m}+{r}_{\mathit{ela}}^{d})\)
(\(\mathit{kPa}\)) |
Code_Aster |
GEFDYN |
Relative error |
15 |
0.326 |
0.326 |
-0.088% |
-15 |
0.395 |
0.394 |
0.172% |
15 |
0.395 |
0.394 |
0.172% |
-15 |
0.579 |
0.578 |
0.255% |
2.4. notes#
The comparison between Code_Aster and GEFDYNest solutions is relatively good. The differences are explained by the integration method chosen for GEFDYNqui is explicit. You would have to increase the load split to obtain values closer to those of Code_Aster.