Explicit meshing - what to do with sharp corners 1

[assafwei]

Hi,

I am modelling crash and drop test for parts with a lot of very small radius features (<1mm). In order to get the run time to reasonable numbers I have to remove these radius features and most of the time I get sharp corners. When using implicit schemes I now this is a big no no, but in explicit I don't really have a choice (a few days run times are out of the question).

How do you tackle this issue? From my understanding sharp corners are less problematic in explicit analysis due to the dynamic nature of these problems, am I wrong?

Thanks.

 

[FEA way]

That’s right, in explicit you can (or even should) remove the small features such as rounds/fillets. It won’t affect the accuracy significantly but will increase the stable time increment and thus speed up the analysis.

 

[assafwei]

Thanks for the input

 

[L_K]

You should only model those radii that really matter (you want to know exact stress in that region).
If even this makes the runtime unreasonable, you could use mass scaling for those small elements.

 

[Mustaine3]

Yes, smart mass scaling with a target time increment can help a lot in such cases.