abaqus mesh error - creep simulation

[Balamugesh M]

i am getting error while submitting the job in abaqus. it tells that "the following part instances are not completely meshed because they contain regions with only boundary elements". what to do for this error? if i check any unmeshed faces it is null. how to rectify the error?

 

[FEA way]

Check the mesh of this instance - type of elements, meshing technique and so on. Try regenerating the mesh and see if there are any errors or warnings. Use the display group Remove Selected tool to hide some elements and see what it looks like inside.

 

[Balamugesh M]

i am getting convergence and elements distortion issue. what to change for this issue?

 

[FEA way]

So the error about part instances not completely meshed is gone now, the analysis starts normally and you just have convergence issues ? Those may have lots of different causes, it's impossible to say without knowing all the necessary details about the model. Are the distorted elements just a warning ?

 

[Balamugesh M]

yes instance problem is solved. now i am getting convergence issue due to elements are distorting. what are the details you needed i can provide it.

 

[FEA way]

Ideally, the whole model file (.cae or .inp). If you can't share it then you should provide as many relevant details as you can - screenshots of the meshed geometry, material properties, step settings, BCs, loads, interactions, constraints and so on.

 

[Balamugesh M]

hi, i cant able to upload the .cae file as it exceeds 20mb i will share you the images. All

in SI unit(mm)

 

[FEA way]

Make sure that the input values (especially creep constants) are correct. The error seems to come from the creep behavior. You could try adjusting the incrementation and CETOL settings as well.

 

[Balamugesh M]

creep constants are correct it worked correctly for a sample rod, if i try with actual product i get this error. is there anything else i need to do?

 

[FEA way]

Then try adjusting CETOL. But it's also important when and how it fails to converge so one should take a close look at the status and message files.

 

[Balamugesh M]

if you don't mind can i ask is CETOL is the tolerance value rt? or anything else? yes we have close look in the message and status files.

 

[FEA way]

Yes, CETOL is defined as "Creep/swelling/viscoelastic strain error tolerance" in Abaqus/CAE. It’s important for the convergence of creep problems.

And what is the conclusion ? Can you confirm thay it fails because of large creep strains ?

 

[Balamugesh M]

thanks for the confirmation. i have kept it to 1e-10 the CETOL value. below image when it starts creep ,step attempt 1

 

[FEA way]

Try increasing CETOL significantly.

 

[Balamugesh M]

ok i will increase it. can you tell me the rough value,

 

[FEA way]

It depends on the problem, it’s best to change it significantly (maybe even by orders of magnitude), try running the analysis and change it if the simulation still fails. For those tests, it might be reasonable to reduce the mesh density a bit.

 

[Balamugesh M]

ok sure, i will try to modify and run it