Multi-support excitation in soil model

[BuuNguyen]

Hi STKO,

1/ I used uniform excitation, it worked well and gave a similar results with the experimental one in terms of acc, disp (after transfering relative to absolute values).
However, regarding horizontal stress, are there any effects of relative or absolute values? or just extract the earth pressure? as in Fig1, horizontal stress increased significantly and kept unchanged even I used much small input motion.

2/ So, Im trying to use multi-support excitation at the base of soil domain, I can get exactly the same input motion at the base, but the acceleration at ground surface is so small (~0).
Can you have a look at my model (multi-support model) and suggest a solution?
Thank for your support,
Buu

[BuuNguyen]

Hi STKO,

I solved my problem with multi-support excitation (second question in previous post) thanks to the note from Opensees wiki: "The non-homogeneous single point constraints require an appropriate choice of constraint handler". firstly I used "Plain constraints" then got some notice abut non-homogeneous point constraints, then changed the constraint handler to "Transformation". It worked well.

However, regarding the horizontal pressure of soil as in Fig 1 above, it increased significantly during EQ and kept unchanged at the end of EQ (residual value) for both uniform and multi-support excitations. I think its not correct

Can you suggest any solution for it? It can be due to boundary condition?

Thank for your time,
Buu

[STKO Team]

solved my problem with multi-support excitation (second question in previous post) thanks to the note from Opensees wiki: "The non-homogeneous single point constraints require an appropriate choice of constraint handler". firstly I used "Plain constraints" then got some notice abut non-homogeneous point constraints, then changed the constraint handler to "Transformation". It worked well.

This is right

However, regarding the horizontal pressure of soil as in Fig 1 above, it increased significantly during EQ and kept unchanged at the end of EQ (residual value) for both uniform and multi-support excitations. I think its not correct

This is normal according to your model.
You fixed the Ux displacement (with fixes during the gravity, and with EDOF during dynamic). It means your soil cannot expand in the X direction.
So you first see a Sigma11 coming from the poisson effect (in the elastic regime).
Then, while your model goes into nonlinear regime during dynamic analsyis, you see the effect of dilation (plastic increase of volume due to plastic shear strain). But since the overall X displacement is fixed, your soil cannot "dilate" and so it will show increasing confining stress

[BuuNguyen]

You fixed the Ux displacement (with fixes during the gravity, and with EDOF during dynamic). It means your soil cannot expand in the X direction.
So you first see a Sigma11 coming from the poisson effect (in the elastic regime).
Then, while your model goes into nonlinear regime during dynamic analsyis, you see the effect of dilation (plastic increase of volume due to plastic shear strain). But since the overall X displacement is fixed, your soil cannot "dilate" and so it will show increasing confining stress

But I already removed Ux constraint before EQ shaking. Now, I changed boundary condition in my model to fix Ux, Uy at base and EqualDOF at 2 sides during both gravity and dynamic. However, the horizontal stress have the same problem as before. you can see it in attached file.

[STKO Team]

What I mean is that in the second stage, even if you remove the Ux fix at the vertical sides, you added an EqualDOF.
Now, imposing Ux_Left = Ux_right means that there cannot be an expansion in the X direction. Since the model has dilatancy, during plastic loading the model would like to expand for both 1) the Poisson effect, and 2) dilatancy induced by plastic shearing.
But since you are imposing an equalDOF in the X direction, the overall (mean) expansion in the X direction must be zero. Therefore, as a reaction to this constraint, you see increasing stress in the X direction.

[BuuNguyen]

It makes sense, but many researchers also used periodic boundary condition to simulate the boundary of centrifuge test.
Do you have any suggestion to simulate the boundary condition of centrifuge test? I have tried AutoSTKOAbsorbing BC, but cannot get the similar acc at free-field surface that I can get exactly the same if I use periodic boundary.
p/s: I have a acc input in the soil at the base of the soil box

[STKO Team]

I think that the periodic conditions are fine for a centrifuge test. Try to google and search for similar issues. For example: "boundary effects in centrifuge tests"