Excavation issue

[[email protected]]

Dear STKO team,

I am trying to use your new zerolength element (implex) for a soil-pile interaction problem. I am trying to excavate some of the top soil before adding the pile, however the integrator seems to be stuck in an infinite loop when I try to remove a model subset. I can see that the code is running and i snot frozen, because I use the multiaxial plasticity soil model that still has its debug messages output during integration. My code would run non-stop for over a day, but nothing happens. No errors, no solution, nothing.

Also, note that I use multiple initial stage analysis on/off commands, because if I try to excavate all soil in one step, upon resetting displacements the code actually can not solve for ~0 displacements. This happens because the soil is in plastic stage during the analysis (this material does not support elastic/plastic switch), but this is acceptable for my analysis.
The infinite loop issue also persists if I only use one initial stage analysis on/off command. It only resolves if I remove all soil in one step. However, in this case I can not add the pile, as the soil displacements can not be set back to 0.

Can you take a look at the model and see why this may be happening?

Thank you in advance for the help!

Bence

[STKO Team]

First, you forgot to remove 3 construction faces in your model. However this is not the source of the issue, since you are not using them in any model subset.

Second, Now you cannot use more than just 1 initial stage ON/OFF pair. I had a look in the source code and they do:

Code: Select all

epsilon_0 = epsilon_0 + current_epsilon

insead of

Code: Select all

epsilon_0 = current_epsilon

upon setting initial stage = OFF.
This means that the second time you call initial stage OFF, it will save in epsilon_0 twice what it had before... It's a bug. I will contact the author and suggest a modification. Until then, you cannot use it multiple times.
But I don't think you need it. You need initial stage off only when you ADD new components, not when you remove them.
The concept is that, when you add a new element, it can't handle an initial displacement as a zero strain, so you need to reset the displacement to 0 in previous elements that will store the initial strain, while the new element will see a zero strain due to zero displ.

Finally, I think there is some issue in the material model. The error message (debug) says that something "strange" happenend in the plastic integrator, and since it is something un-expected, it might be that the code never lives the plastic integrator loop. Try to change the material

[[email protected]]

Dear STKO Team,

I have talked to Prof. Wang Gang, the developer of the soil model I am using, and trying to validate (hence, I can not change the soil material for this problem) and he said all the messages are irrelevant and do not indicate issues with the code. It is simply debugging code that requires some clean up.
I have managed to excavate the soil for my pile by placing the pile in before excavation (it is crude and leads to some inaccuracies, but I can afford some errors in this case). However, the implex element does not show any gapping behavior during the pushover analysis. Soil-pile de-bonding never occurs.

Result.png (257.28 KiB) Viewed 2289 times

Moreover, I tried the correct construction steps (excavation prior to pile installment) however, in this case after issuing the "initial stage OFF" command, the soil displacements were not reset to zero (in the order of dm persisted), and there was an overlap between the pile and the soil during the pushover analysis.

Do you have any suggestions on how either of these issues may be solved?

I have attached the STKO model file. The only difference between the two discussed models is the location of the analysis step, "Add pile" (1.: together with "Add soil" 2.: after the "@@@dummy analysis").

Thank you for your help!

Bence

[STKO Team]

Dear Bence,

Unfortunately checking this model will take some time.
We have updated the support conditions (see Maintenance Packages at https://asdeasoft.net/?stko-purchase) and the support for reviewing large complex models should be requested.

So I would suggest you create a simplified version of this model, using simple linear elastic materials for the soil so that we can check the behavior of the Contact element in a reasonable amount of time, and this is included in the Siver Maintenance Packages you have.
Otherwise, it's difficult to understand where the problem is.

[allano89]

Dear Bence
I´m trying to use the multiaxial cyclic plasticity but I didn´t find how it works. Maybe do you have some references about how to calculate the parametrs Ho y h?
Please your help
Best regards
allano89