excessively distortion and too large simulation time

[jingansong]

Hi everyone,

I'm trying to do a simulation that a sphere slides on a plate, and meanwhile it moves toward the plate surface (like penetrating into the plate). However, it always fails at the beginning because of excessively distortion. After I improved the mesh quality, it can process further, until the excessively distottion problem happens again. I double checked the material properties difinition, and the unit consistance, it looks correct.

Also, according to what has finished, the simulation may take two months to complete on a 12 cpu cluster. Previously, I did some other simulations on the same cluster, the sizes of the input files of those simulations were larger than the size of the current one, but they only needed one week to complete. The difference is that the current simulation is smaller structure based on nanometers, while the previous simulations were based on micrometers. In abaqus, I use micrometer based units for both the current and previous simulations.

Could you please help me take a look at my model definition and give me some possible reasons and solutions? Thank you very much!

https://www.dropbox.com/s/7kb8bcsdsepiunm/sp-20130503.cae

 

[IceBreakerSours]

I did not look at your model but have you tried using adaptive remeshing? From your description, it sounds like remeshing may just be the solution for you.

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[jingansong]

Thank you IceBreakerSours. Do you mean adaptive mesh? I did use adaptive mesh in the simulation.

 

[IceBreakerSours]

What warning/error messages do you see in the .dat/.msg file?

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[jingansong]

Hi IceBreakerSours. There are some warnings related to output request. And one warning is "Some nodes in the general contact domain are part of a sliding or eulerian boundary region. These nodes will be made lagrangian". But I saw this one before in previous simulations, I don't think this is the reason.

I saw that the stable time increment for some elements is very small. But these elements are just cubes rather than poorly meshed elements. I'm wondering that what causes this may be the reason for all my problems.