Hello Code_Aster community,
I am currently performing a dynamic soil–pile interaction analysis in Salome-Meca 2024 (WSL) using DYNA_NON_LINE. The model consists of a 3D soil block with a pile embedded in it. The soil is modeled using Mohr–Coulomb plasticity, while the pile is modeled as elastic.
The analysis procedure is as follows:
- Static step (STAT_NON_LINE) from t = -1 to 0 to establish the gravity equilibrium state.
- Dynamic step (DYNA_NON_LINE) from t = 0 to 200 s where seismic loading is applied.
For the seismic excitation, I impose a horizontal base displacement at the soil base using a sinusoidal function:
u(t) = A · sin(2πft)
where:
- f = 10 Hz
- A is the displacement amplitude derived from acceleration values (50, 100, 200, 400 gal).
The displacement amplitudes were computed using:
u₀ = a₀ / ( (2πf)² )
So the cases correspond approximately to:
- 50 gal → 1.27E-4 m
- 100 gal → 2.53E-4 m
- 200 gal → 5.06E-4 m
- 400 gal → 1.01E-3 m
However, when I run separate simulations with these different amplitudes, I obtain almost identical response curves at the pile head (displacement and reaction). I expected the response to scale with the input motion amplitude.
Key modeling details:
Soil: Mohr–Coulomb with PETIT deformation
Pile: Elastic
Rayleigh damping implemented with AMOR_RAYL_RIGI='ELASTIQUE'
Time integration: HHT scheme with ALPHA = -0.2
Gravity applied during static step and maintained during dynamic step
Boundary conditions:
base constrained in DY and DZ
horizontal displacement applied at the base in DX
lateral boundaries constrained in normal directions
I have already checked the following:
- Different .comm files were used for each amplitude case.
- The displacement amplitude in the sinusoidal function was correctly modified for each run.
- The dynamic step starts from the static equilibrium state (ETAT_INIT = EVOL_NOLI).
Despite these checks, the results remain almost identical between cases.
My questions are:
- Is the base displacement loading method appropriate for this type of dynamic SSI problem, or should I instead apply base acceleration using CALC_CHAR_SEISME?
- Are there issues in the boundary conditions or excitation definition that could cause the response to remain unchanged with different amplitudes?
- Could Rayleigh damping or solver settings be masking the amplitude differences?
- Are there recommended practices for dynamic soil–pile simulations in Code_Aster that I may be missing?
I would appreciate any suggestions on how to improve the .comm file or the loading strategy to obtain physically meaningful dynamic responses.
Thank you very much for your help.
