Nastran sol 129 non-linear transient analysis - Gap element usage

[MOZER8]

Hello All,

I'm new to Nastran and have a question related to SOL 129 solver.

I have a model which is running perfectly in SOL 109 (linear transient) solver. For some reason I needed to use gap elements for contact problem, thus I switched to SOL129. However gap elements are not working properly. They move to second stifness section before the load is applied and do not return to their original position even though there is no longer loading in the system. Could you help me to understand what's wrong with the model?

An example model was built and attached in .bdf format. Since it is not that big, it takes several seconds to solve in moderate computer. Thanks in advance for your support. Let me know if you need any additional information about model.

Regards,
Mehmet

 

[Erik Panos Kostson]

Not an expert in Nastran - I only know because in the software I use we can import .bdf files.

Not used to the initial gap, but I think you set it to 22 then it needs to travel the whole element length+2 mm in order to become active which is strange.

I would set it to 0, and then it (cgap) should be active as soon as it is in compression (that is the distance between its nodes is getting smaller).
Not sure also if the PID of 2 (cgap def.), is missing on the PGAP definition, because the first thing seen is 22.0 there, which I assume is the initial gap since it is a decimal value.

 

[MOZER8]

Thank you for the prompt answer Erik.

After you mentioned I checked the .bdf file, because I remember that the initial gap was set to 2mm.

Nastran uses 8 digit system in bdf input files. Thus it is not 22.0, it is 2 and 2.0. First 2 indicates to pgap ID and second 2.0 indicates initial gap. Usually those values are seperated with space however hypermesh exported files sometimes have these bad format issues(model was built in hypermesh). There is no error in format but it is hard to read.

 

[dfreire]

Hello,

If you are using Altair software, it is better to get the results as *.op2 file (see attached image). Hyperview does not like very much the *.xdb files. You only need to change the POST parameter from 0 to -1, and you can also remove the DBFACT=4, it will be no needed.

I would follow the next plan:

- As QRG says, in remark 8 of CGAP entry :"8. Since a large stiffness is used for KA (the closed GAP stiffness), param,g damping should be avoided. Instead damping should be specified on the MATi entries and PARAM,W4 set." I would try this, it is very quickly and you can see what happens.

- Maybe the loading is not enough to close the gap completely. You can make a easy check by swiching the CGAP elements for CELAS1 or CBUSH with an axial stiffness of 10.0N/mm, that is the defined for the open gap. If you get the same result, it means that the applied load will not close the gap.

I hope it helps.

 

[ThomasH]

"Nastran uses 8 digit system in bdf input files".

That is not entirely correct. That applies to the "Small Field" format. You can use "Large Field" and get (I think) 16 digits.

But that has probably nothing to do with your issues .

Thomas

 

[MOZER8]

dfreire,

-Thank you for the suggestion. First of all, I changed the output request to .op2. This solved the missing element results issue.

-As you mentioned, I removed G damping used MATi and PARAM,W4 instead. While defining G damping, my intension was to provide stability to gap elements. Because the problem is not only incorrent gap during the analysis, but also irrational gap element behavior. For better explanation, I attached a picture which shows the loading during the analysis. Although 200N ramped force begin to act on model in t=0.1s, gap elements starts to move before t=0.1s. Afterwards upper solid block continue to move independently from the force. I also upload updated version of .bdf file, you can observe the motion that I tried to explain.

-I did hand calculation, initial gap should be close. I also tried different force and stiffness values, however it seems like gap element behave like it doesn't affected by force and stiffness as I mentioned above.

I don't know if I'm missing something.

BDF:

https://files.engineering.com/getfi...c6c9&file=Nastran_gap_nonlinear_transient.bdf

Loading:

https://files.engineering.com/getfi...-9ab1-48c8-83ac-9f71b76db296&file=Loading.png

Regards,
Mehmet