Meshing very thinly spaced regions in ABAQUS with error ErrElemAreaSmallNegZero?

[JAIN-FEM]

ERROR WHILE RUNNING THE JOB
Error in job : The area of 12 elements is zero, small, or negative. Check coordinates or node numbering, or modify the mesh seed. The elements have been identified in element set ErrElemAreaSmallNegZero.

DESCRIPTION
When checked in the JOB NAME-->MONITOR-->DATA FILE, it is because the element size of few elements is too small. My problem is that i can't compromise with the geometry. The region where these erroneous elements lie is located between two partitions which are closely spaced, which can not be spaced any further apart. Could you please suggest any workaround?

Thanks

 

[Mustaine3]

Depending on what you use you can try a different meshing technique, maybe deactivate mapped meshing.
Otherwise you have no choice than reducing the element size in that region to get a better element shape.

 

[FEA way]

Can you attach a picture of this region ? Seems that you will have to apply some local seeds there.

 

[ModuliTechnologies]

At times Abaqus solver needs a very tiny clearance (gap) between two contacting surfaces. This doesn't mean anything in real life but helps the solver to start computation using the contact pair type of interaction (assuming you are using CP). You mentioned that you can't space them further apart, but any gap created (also in the order of 1e-3 mm gap) will help the solver, in my opinion.

Moduli Technologies

http://www.modulitech.com

 

[Mustaine3]

At times Abaqus solver needs a very tiny clearance (gap) between two contacting surfaces. This doesn't mean anything in real life but helps the solver to start computation using the contact pair type of interaction (assuming you are using CP). You mentioned that you can't space them further apart, but any gap created (also in the order of 1e-3 mm gap) will help the solver, in my opinion.

First of all I think, that your recommendation has nothing to do with the issue of the TO.

But more important - you are very wrong. The opposite is the case. You should avoid gaps (and overclosures) as much as possible between contact region. Especially when running an analysis force-controlled. Abaqus even has options to achieve that more easily. Think about the adjust-option for contact pairs or the contact initialization option for general contact.

 

[DanStro]

Without seeing an image it is impossible to tell if my comment will be relevant but here it is anyway:

I have had similar problems with thin parts and contact initialization. Sometimes the adjustment that is done creates elements with negative thickness. In my case this is can often be corrected by tweaking the model but if you can't do that then if you have contact initialization set for those surfaces try removing them.