Modelling soil liquefaction in linear elastic

[aboufira]

Hello,

I am just getting back into OpenSees, but for geotech this time. I would like to simulate the following problem:

- Geometry: Soil Layer 20 meters wide and 10 meters deep
- Loading: -1 kPa on top of the soil layer going from x=0 to x = 0.5. Loading remains constant throughout time domain.
- Constitutive Model: Linear Elastic, Undrained
- Type of Analysis: Dynamic Effective Stress Analysis
- Boundary conditions: Fixed in normal direction at x =0 and x = 20. Fully fixed at y = 0.
- Type of Output: vertical effective stress, pore pressure (excess), displacement, velocity, acceleration.

I also have a sketch of the problem but it seems like there is no way to attach images...

I have not found material models that would allow for this sort of simplicity. What (I believe) is possible comes from the following link (http://soilquake.net/opensees/) which uses mainly PressureDependMultiYield coupled with FluidSolidPorousMaterial (Example 15). Is there any way to remove all nonlinearity/plasticity from the model and still investigate solid and liquid phase dynamic response in OpenSees?

[aboufira]

Allright. So I have succeeded in generating a soil model as shown in the following image (https://postimg.org/image/cat08noo1/) using GiDOpenSees interface. I am postponing the liquefaction part for later and just focusing on the dry material.

- General Problem Information: 2DOF and 2dimensions.
- Element Type: Quad elements in Plane Strain
- Material: nDMaterial ElasticIsotropic

The loading and analysis are as follows. There will be no initialization of stresses due to gravity. I am only interested in the dynamic output and do not wish to overcomplicate my model.

############################
pattern Plain 100 Constant {
load 215 0.000 -1.000
load 222 0.000 -1.000
}

# Analysis options
test EnergyIncr 1.0e-12 10 0
system BandGeneral
numberer RCM
constraints Plain
integrator Newmark 0.5 0.25
algorithm Newton
analysis Transient
analyze 100 0.01
#####################################

As mentioned before this is a transient dynamic analysis but with constant loading throughout the time domain. The analysis appears to run smoothly but unfortunately when I monitor the displacement at node (X=0, Y=0.5) I am getting a constant displacement of 0 (straight line) throughout the time domain. If I switch the loading to linear I get a linear displacement.

I know from other software this should not be the case for dynamic analysis. Even if the load is constant, the fact that it is suddenly applied at t=0 causes a 2D wave to propagate throughout the medium. This should be reflected in the displacement at X=0 and Y=-5.

Could anyone please provide insight as to what I am doing wrong here?