SSI Simple Model Problem

Post Reply
User avatar
vokumus
Posts: 19
Joined: Thu Nov 14, 2019 7:07 pm

SSI Simple Model Problem

Post by vokumus » Sun Nov 17, 2019 7:41 pm

Hi;

I am trying to create a very simple SSI model for my PhD thesis. I have read many papers and watched videos of STKO and created a model. But after 2 days of hard work I couldn't manage to get acceptable results. Displacements and other values are just astronomical.

Model file is attached. I appreciate any kind of help.
Attachments
c03-elastic-quad.zip
(136.96 KiB) Downloaded 31 times
Last edited by vokumus on Tue Nov 19, 2019 11:08 am, edited 1 time in total.
PhD candidate. Licenced STKO user. C# Developer.

STKO Team
Posts: 71
Joined: Tue Oct 29, 2019 8:45 am

Re: SSI Simple Model Problem

Post by STKO Team » Mon Nov 18, 2019 9:42 am

Hi,

Extreamely large displacements are due to the fact that you swapped the values for gamma and beta in the newmark method.
Another problem: the alpha parameter in the L-K boundaries should be 1 for linear damping. With alpha = 0, the stress given by those dashpots will always be = 1.

Other than this, we didn't check that all other parameters are correct or consistent with the unit system you are using.

Let us know if this solves your problems

User avatar
vokumus
Posts: 19
Joined: Thu Nov 14, 2019 7:07 pm

Re: SSI Simple Model Problem

Post by vokumus » Tue Nov 19, 2019 6:31 am

Changing Newmark parameters with correct values solved most of my problems. Don't know it is appropriate to ask here but I have some questions that I couldn't find clear answers:

Wilson says in his "3d Analysis of Structures" book;

"To activate the soil structure interaction within a computer program, it is only
necessary to identify the foundation mass so that the loading is not applied to
that part of the structure."

Then he says:

"Most general purpose programs for earthquake analysis of structures do not have
the option of identifying the foundation mass as a separate type of mass on
which the earthquake forces do not act. Therefore, an approximation that has
commonly been used is to neglect the mass of the foundation completely in the
analysis."

Then he finds similar results with mass included and mass neglected models.

So, what is the best way to model it in OpenSees. Should I neglect soil mass in the analysis?

By the way, STKO is such a great software. Congratulations with your team's great work.
PhD candidate. Licenced STKO user. C# Developer.

STKO Team
Posts: 71
Joined: Tue Oct 29, 2019 8:45 am

Re: SSI Simple Model Problem

Post by STKO Team » Tue Nov 19, 2019 10:09 am

We are glad you are enjoying STKO!

I will have a look at those sentences as soon as possible... At a first look, it's not clear to me what was the point there... I'm sure some experts in geotech and SSI will reply soon to this post.

User avatar
vokumus
Posts: 19
Joined: Thu Nov 14, 2019 7:07 pm

Re: SSI Simple Model Problem

Post by vokumus » Tue Nov 19, 2019 11:07 am

Thank you very much. I will talk with my thesis supervisor soon and share his ideas here.
PhD candidate. Licenced STKO user. C# Developer.

Jenny
Posts: 2
Joined: Mon Jul 22, 2019 1:00 am

Re: SSI Simple Model Problem

Post by Jenny » Sun Nov 24, 2019 6:05 pm

Hi
I think you should always consider the mass/density of the soil. Note I am assuming that you are modeling the soil with cube or quad elements. Currently, I cannot open your file with the STKO version I have, but I'll try to fix the problem shortly. In the meantime, I will add a couple of thoughts regarding how to model foundations in opensees. In order to model a shallow foundation, some people use (i) beam elements and springs, (ii) solid elements when using 3D models, (iii) quad elements when using 2D models, (iv) and others. For case (i) the mass is usually added manually on every node. In case (ii) or (iii) the density of the element times de gravity is defined in the element. Sometimes, this depends on the constitutive model is associated to your element to understand how the mass is applied, Hopefully, you are not double counting effects that are producing large displacements. This is a short explanation, but I would be happy to check your model (once I can) and extend these thoughts based on your model's conditions. Regards.

Yu-WeiHwang
Posts: 4
Joined: Sun Nov 24, 2019 3:30 am

Re: SSI Simple Model Problem

Post by Yu-WeiHwang » Sun Nov 24, 2019 7:20 pm

Hello,

The mass of the soil will affect how the wave propagate, so typically we will want to consider the mass of the soil to better capture the seismic loading. Another thing you may want to check is whether your designed fundamental frequency of the structure included the foundation or not. Sometimes when you calculate the fixed based frequency, they don't consider the foundation in the model or physical testing. On the other hand, the foundation will affect your stress distribution a lot when you do the modelling. Depending on what kind of SSI problem you are interested in, you can consider different way to approach it.

It looks like you used a very large thickness for your quad plane-strain element. You may want to check it is that reasonable or not. When I run SSI in 2D, I only assign 1 to the thickness of the soil and foundation element.

User avatar
vokumus
Posts: 19
Joined: Thu Nov 14, 2019 7:07 pm

Re: SSI Simple Model Problem

Post by vokumus » Tue Nov 26, 2019 11:03 am

@Jenny and @Yu thanks for your replies.

I want to investigate SSI behaviour of a high-rise building with a pile raft foundation for my thesis. Models will be fully nonlinear and 3d but my supervisor suggested me to work on simple 2D and elastic models first.

There are many papers and thesis about SSI but I couldn't find a clear answer whether or not to consider soil mass. So that's why I have asked it here.
It looks like you used a very large thickness for your quad plane-strain element. You may want to check it is that reasonable or not.
At the beginning I decided to work with kN/mm units which is a bad decision. I now have changed my unit system to kN/m.

This week I'll talk with my supervisor. I'll share his thoughts and my model here soon. Thanks in advance.
PhD candidate. Licenced STKO user. C# Developer.

User avatar
vokumus
Posts: 19
Joined: Thu Nov 14, 2019 7:07 pm

Re: SSI Simple Model Problem

Post by vokumus » Tue Dec 10, 2019 10:24 pm

Hi again.

We have discussed the mass issue with my thesis supervisor Prof. Yasin Fahjan. He said that If I neglect soil mass, it's meaningless to do SSI analysis. For kinematic interaction soil mass have to be considered.

Best regards.
PhD candidate. Licenced STKO user. C# Developer.

Post Reply