Mass Scaling issue in ABAQUS/Explicit

[ShadowWarrior]

I have a drop impact test problem, that I'm simulating in ABAQUS/Explicit. The mass of Deformable body is 0.3693672 gram and Rigid flat impactor plate is 272 gram (which is dropping on the deformable body at 2.8 m/s velocity). Mass scaling is being applied to the deformable part only.

Now -

Case 1: If I limit the % mass increase (DMASS) in whole model to 5%, the analysis finishes in 35 hours on a high performance computing system. But DMASS in the deformable body becomes 5,000%! Bringing the final mass of deformable body 18.46836 gram.

Case 2: If I limit DMASS in the deformable body to 5%, DMASS of whole model stays close to zero and it will take ~120 hours to finish on the same computing system.

I know that DMASS < 5% rule is applicable to part level (deformable body only), but can I go with Case 1 for my analysis (minimum run time), considering the deformable body is 736.4 times lighter than the impactor?

Thanks!

 

[Mustaine3]

Why don't you apply the Mass Scaling to the deformable part only? With energy output you could do the same.

 

[ShadowWarrior]

Mass scaling is being applied to the deformable body only, I have edited that to the original post. However, even if I applied Mass scaling to whole model, it will be same because there is one deformable body and one rigid body in the model and mass scaling is not applicable to rigid body.

Whole model DMASS takes into account the mass of both the deformable body and impactor, whereas deformable body only DMASS takes only the mass of deformable body into account. So if I limit whole model DMASS to 5%, the deformable body DMASS becomes ~5000%. If I limit deformable body only DMASS to 5%, it will take ~7 to 10 days to finish the analysis.

So what should I do?

 

[IceBreakerSours]

Can you try element based mass scaling and see how that works out?

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

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

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

 

[ShadowWarrior]

Element based mass scaling? If you mean variable mass scaling, I'm using that.

From my input file - *Variable Mass Scaling, elset=deformable_body, dt=5e-09, type=below min, frequency=3

There is no error in Mass scaling approach. Problem is excessive mass increment if I try to speed up my analysis, Or Excessive computing time requirement if I try to stick to <5% DMASS rule in deformable part only.

 

[Dave442]

Hi,

Can you justify the 5,000 fold increase in mass? What does this do to the energy balance? Compare the results and energy balance for the 35 hrs analysis to those from the 120 hrs analysis. Is there much difference? If so, is it significant? Do the results still line up with what you see in reality? It's up to you to justify the quicker approach, you know more about the problem than anyone here. Otherwise, your stuck with 120 hrs run time or de-featuring/re-meshing to increase the size of your critical elements.

Good luck,
Dave

 

[ShadowWarrior]

Hi Dave,

Most of the inertia effect will be dominated by the mass of the impactor, since its 736.4 times heavier than the deformable body. So my guess is the large increment in mass of the deformable body would not affect that much on the overall dynamic response, but its just a guess. I think I should run the 120 hours analysis once and then compare it to the 35 hours run, as per your suggestion.

Now the question is, if I see 35 hours run is similar to 120 hours run, how do I present/defend this unusual mass increment approach? Or should I just omit it from mentioning?