High Displacements with Plastic Material Stage

[STKO Team]

Dear Vokumus,

We had a look at the file and you were right, we put slightly wrong values. I corrected them and I noticed that the response gived by using only elastic materials (i,.e. without switching to plastic) are acceptable but not very good. While in the plastic regine they are wrong.

First of all keep in mind that the dashpot coefficients are computed based on the shear modulus, so if they remain constant during the analysis, it will provide a good prediction only if also the material remains elastic. Unfortunately the dashpot coefficient cannot be updated during the analysis. It can be done if you have some c++ coding experience with OpenSees.

[vokumus]

Dear STKO Team;

I see. I'll try to work on the dashpot issue soon.

Thanks for your support.

[vokumus]

Hi;

I talked with one of the professors, Prof. Yasin FAHJAN at Istanbul Technical University, and we concluded that main reason of these lateral deformations is the soil's own weight. If soil weight is not applied, problem disappears.

We think this is because soil weight causes a lateral pressure. Prof. Ahmed Elgamal has a note for this issue in OpenSees PL software manual:

"Apply own weight of the soil using a global lateral stress coefficient, and a single value of Young’s modulus that is user defined (this will reduce initial shear stresses in the mesh due to own weight application, but generally will have minimal impact on the subsequent earthquake computations anyway),"

So you can use a K0 value and decrease the soil weight (does not completely solves the problem but results become better), or neglect the soil weight completely if you have interested in the dynamic behavior only.

I decided to write this here so other users having the same problem can benefit. Hope this solves the mystery

[STKO Team]

Thanks vokumus, if it is a known issue it is worth working on it. If you conduct more tests on it, please post your results here, it will be interesting.

[vokumus]

Ah yes, we are doing some tests with my thesis supervisor and I will share them here soon.

Best regards.

[STKO Team]

Thanks!

[vokumus]

Update:

I talked with Prof. Pedro Arduino this weekend and he suggested this method If you want to get rid of this lateral displacement accumulation. You can create a thicker (10~100 times thicker) region at the end of the boundary, just before the free field columns(Also the free field columns). I tried this and results are quite nice.

This method is used in his official example too:

https://opensees.berkeley.edu/wiki/inde ... s_of_Slope

Also, during my tests, I did not encounter this problem in 3D analyses. I think this issue is specific to 2D analysis.

Regards.

01.jpg (224.16 KiB) Viewed 4571 times

02.jpg (74.75 KiB) Viewed 4571 times

03.jpg (81.62 KiB) Viewed 4571 times

[STKO Team]

Yes, you are right,
when we did some tests in 2D we used that increased thickness too. Unfortunately, our only webinar on this topic was in 3D where this effect was not visible, probably because the FF layer in one direction is coupled with the other direction through the corner.

However, increasing the thickness in 3D is not possible, so at most, you can use increased elastic properties. It would be nice to understand why one needs such a huge increase in thickness only in 2D.

You can do a test.
Use a model where the Width of the FF is L, and the thickness is 1.
Use a model where the Width of the FF is L, and the thickness is 10. (increased thickness)
Use a model where the Width of the FF is 10*L, and the thickness is 1. (increased width)
You can see if the results of the 2nd and 3rd options are the same.

[vokumus]

I'll look at it as soon as I get the chance. Thanks.

[Horace Horsecollar]

Great!