Ground loss simulation

[Lei Lei]

Dear STKOteam,

The objective is to simulate a scenario of ground loss, where an initial gap exists between the tunnel and the soil.

6d8b20b48d1bdbc25beaae7a5f06722.png (40.08 KiB) Viewed 3482 times

When gravity is applied to the soil, soil deformation eliminates these gaps and forces contact with the structure. Additionally, due to soil characteristics, the interaction behavior between the soil and structure is purely compressive rather than tensile.

Is there an interactive behavior or strategy that simulates this type of contact?

I've already tried to use ZeroLengthImplexContact Element, while the results appeared to be anomalous.

f8459a877b6c0f7e09f8d7d51fdef59.png (107.38 KiB) Viewed 3482 times

Looking forward to your reply!

[kesavapraba]

Hi, you can try zerolengthContactASDImplex. Please refer to the following links:

GitHUB page about the contact element:
https://opensees.github.io/OpenSeesDocu ... mplex.html

Dr. Massimo's webinar about the contact element:
https://www.youtube.com/watch?v=tXA5guozrFU

[Lei Lei]

Dear STKOteam,

Thanks for your advice. I've tried the zerolengthContactASDImplex element to simulate the behavior of the contact with a gap between the soil and tunnel, which has already removed the section extrusion of the tunnel. Just like figure below：

6d8b20b48d1bdbc25beaae7a5f06722.png (40.08 KiB) Viewed 3450 times

Here, I clicked the rigid link. Otherwise, it won't compute the excavation stage(the third stage). The error message is as follows.

6cee07568e24bd4634d322f2f2754f9.png (81.09 KiB) Viewed 3294 times

However, the rigid link can't simulate the convergence of the gap between soil and tunnel. The results is below

7997122975409d9656dc0574140019d.png (279.05 KiB) Viewed 3294 times

So, is there any other contact element to simulate the gap convergence?

And, how to choose the appropriate value of the kn and kt？

Looking forward to your reply！

[kesavapraba]

Hi, It follows Mohr-Coulomb frictional law. Please note, Kn and Kt define linear limits (beyond that slip occurs)
Kn = Penalty factor for normal action = This can be 2 to 3 order higher than "E" value
Kt = Penalty factor for tangential action = This can be about G times the thickness of the contact interface OR one order higher than the G value

Please check out the below figure for more understanding. Also, the link below

https://opensees.github.io/OpenSeesDocu ... mplex.html

[Lei Lei]

Thanks for your reply!

one order higher than the G value

I think this may be the most useful method.

thickness of the contact interface

By the way, isn't the thickness of the contact interface zero?

And, since the penalty method can simulate the behavior of separation of the connected two elements, can it simulate a gap closure before they connect?

I tried to use the zerolengthContactASDImplex element to simulate the behavior of the contact between soil and tunnel with a gap. I clicked the rigid link, and the gap was undeformed. If I didn't click it, the computation wouldn't go on.

Can't STKO simulate this behavior in which soil fills the gap under gravity like what fluid does?

Looking forward to your reply!

[kesavapraba]

Hi,

By the way, isn't the thickness of the contact interface zero?

Yes, in this case. However, in general, you can also have non-zero thickness interface. Refer to some books on Contact Mechanics. In case you want to avoid modeling the geometry of the inteface by assuming it the contact bodies are very nearby, you can multiply the thickness with G to arrive at the parameter that we discussed above.

I tried to use the zerolengthContactASDImplex element to simulate the behavior of the contact between soil and tunnel with a gap. I clicked the rigid link, and the gap was undeformed. If I didn't click it, the computation wouldn't go on.

In order to have gap that close or open, you should not enable the rigid link option. Please note that rigid link is used to maintain the physical gap as it is during the analysis, no closure possible! If your analysis fails without rigid link, there may be other issue with your model definition. Please check it again.

Can't STKO simulate this behavior in which soil fills the gap under gravity like what fluid does?

STKO offers powerful features with which one can model and view results. But, the analysis takes place in backend using OpenSees framework. So, the behaviour simulation got to do with OpenSees. This behaviour is related to fluid dynamics. OpenSees framework has the capability to address this simulation by PFEM approach. You may take a look at the following link.
https://openseespydoc.readthedocs.io/en ... icmds.html

[STKO Team]

Note that the rigidGap option in the contact element is used to simulate the "rigid" half-part of the cross-section of the beam. Otherwise the section will be seen as a Void (Gap).
If I understood correctly your geometrical gap is made of 2 portions: 1) is the radius of the beam-cross-section, 2) the other is the physical gap from the section boundary to the soil. In this case it's better to model the tunnel as a solid and forget about the rigidGap option (used for beams)

[Lei Lei]

If I understood correctly your geometrical gap is made of 2 portions: 1) is the radius of the beam-cross-section, 2) the other is the physical gap from the section boundary to the soil. In this case it's better to model the tunnel as a solid and forget about the rigidGap option (used for beams)

Yes, you are right. So, If I use a solid to model the tunnel with a physical gap from the solid boundary to the soil, can the zerolengthContactASDImplex element model the compression-no tension behavior of the gap without clicking the rigidGap option?

Best Wishes

[STKO Team]

Exactly, with a solid model and a physical gap and no -rigidGap, the contact will activate only when the gap is closed, i.e. when the structure goes in contact with the soil