asdeasoft.net

Dear STKO,

I am trying to simulate a moving load in a saturated medium. Since the source is inside the model, I have to use the distributedLK3D function in the lateral boundaries to model the absorbing boundaries. In addition, in order to reduce the number of nodes, I have constrained kinematically different boxes by using the Embedded Node to Element function.

In order to reset my displacements, I use the InitialStateAnalysis On/Off with the Initial State Wrapper. After I perfom a dummy analysis to mitigate any vibrations. Then, I remove the Lateral fixities, replace them by equivalent reaction forces and activating the L-K condition. Then I apply my moving load function and I run the transient analysis.

However, the equivalent reaction forces are not applied correctly and displacements are observed in the model. I have tried to run the analysis by keeping the lateral fixities on and the simulation is running as expected, but I want to be able to use the L-K dashpots.

In the attached model, I use the same discretization for all the boxes in order to have the model running fast enough. Could you please have a look at the model and help me fix this issue?

In addition, when defining the properties of the DistributedLK3D should I multiply the Vs*rho with the element size in the face? That's what I have done in the attached model (element size=2m)(I have tried with 1m too and still the same issue). I am working in kN, m, s.
Kind regards,
Konstantinos

In addition, when defining the properties of the DistributedLK3D should I multiply the Vs*rho with the element size in the face? That's what I have done in the attached model (element size=2m)(I have tried with 1m too and still the same issue). I am working in kN, m, s.

The distributedLK condition already multiplies by the tributary area (that's why "distributed").
Make sure the dashpot coefficient matches the inner material properties.
If you want to use a more automatic approach, you can try the ASDAbsorbingBoundary element.

Many thanks for the reply.

The ASDAbsorbingBoundary element seems to be recommended primarily when excitation originates from outside the model. Additionally, if I attempt to apply absorbing boundaries on only one side of the model, this element doesn’t work.

I’ve set up a simpler model with parameters: E = 1000 kPa, v = 0, ρ = 2 MG/m3, resulting to Cp = 44,721 and Cs=31,623
Despite removing the lateral fixities and applying equivalent reaction forces, I observe lateral and vertical deformations in the transient analysis. Could you take a look at the model? On a side note, I tried one model where there were only 3 DOF (Without pore pressure) and the model was working fine.

Kind regards,
Konstantinos

Yes it's due to the pressure that change during the analysis and leads to that expansion in the bottom part (the forces from reaction are set from the previous stage, they are not aware of the change in the pressure)

Hello,

Shouldn't the reaction forces already include the horizontal component of the hydrostatic pressure from the previous phases? If yes, there is no reason for a loss of equilibrium. If not, then it makes sense to have this expansion in the bottom part. Do you know perhaps what is the possible solution?

I tried to impose the horizontal component of the hydrostatic pressure on the lateral faces as a function of z-axis, but now the equilibrium is lost on the edges between top and bottom face.

Thanks.

I fixed the issue. Just for the information of other users, I am using the SSPBrickU-P element. In this element, the water pressures are calculated by performing a transient gravity analysis and the water pressures are not included in the reaction force, but in the reaction force included dynamic effects.

In order to achieve this, I specified in the analysis_step.tcl file, before the nodeReaction command, the reactions -dynamic.

reactions -dynamic
load 1 [nodeReaction 1 1] [nodeReaction 1 2] 0.0 0.0
load 2 [nodeReaction 2 1] [nodeReaction 2 2] 0.0 0.0
load 3 [nodeReaction 3 1] [nodeReaction 3 2] 0.0 0.0

Great