Femap - meshing

[feaplastic]

Hello,

I have part with 600*300*10 mm. I need atleast 3 layers of elements thru 10mm thickness and same time less no of elements on other surfaces. Global meshing produces very fine mesh on every area.

What should be my approach? I am new to femap.

Thanks,

feaplastic.

 

[trainguy]

Why don't you use plate elements at the mid-surface of the part?

tg

 

[GBor]

As trainguy suggests, I'm not sure why you are trying to brick mesh a part this thin, but if that is a requirement, develop the mid-plane (or either outer surface) with a tolerable mesh and copy it to the outer surfaces (you may end up with a questionable aspect ratio, but it may work).

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group
Magnitude The Finite Element Analysis Magazine for the Engineering Community

 

[feaplastic]

Hello,

I have to use hyperelastic material with large deformation. So, Plate element is not usable. My part is pretty complex with ups and downs than i explained in my OP. So, It is not possible to copy from outer or midsurface.

One method i found was, mesh all outer surfaces with polt only elements as required and create solid mesh using plot only elements. It worked fine for simple model. Even it has mesh propagation density control. Is it a right method or Any other idea?

Thanks for your replies.

feaplastic

 

[ThomasH]

The meshing as such should not be a problem. You simply map the surfaces and then mesh the volume/solid.

But 10mm thick and 3 elements will will give you a very high aspect ratio och a huge model considering the other dimensions. I would try to convert the 3d material model to a 2d model. If one dimension is much smaller tanh the other you usually have a plate.

Good Luck

Thomas

 

[GBor]

Which software are you using specifically (NE, NX, MSC, ???)? Hyperelastic with large deformation are material models and LaGrangian formulations. Shell elements can handle both in most FEA pacakges and there are methods for dealing with the degree of freedom issues between shells and bricks, if you need to mix the element types.

 

[feaplastic]

Gbor,

I use Abaqus and already tried to model with shell element. It gives a error msg straightaway.

Thanks,

feaplastic

 

[GBor]

Hmmm...we're in the Nastran forum, but what kind of error message did Abaqus give you?

 

[feaplastic]

Gbor,


sorry for the confusion. Preprocessor is femap and analysis is in abaqus. Abaqus says i could not use shell element for hayperelastic model. Is it possible in Nastran?

Thanks,

feaplastic

 

[GBor]

It may depend on which "flavor" of nastran. Pretty sure NEiNastran will do it...haven't used hyperelastic material model specifically with shell elements, but I have used shell elements for extremely large deformation. Not familiar with the specifics of NX or MSC.

I'm sure there are others prepared to give an opinion that know a great deal more about the specifics of other packages.

 

[brep]

feaplastic,

What version of Abaqus, what shell element type, and what material model were you using, and what error message did you get?

I'm not seeing any errors in my test here (Abaqus V6.6, S4R elements, Neo Hooke hyperelastic material based on test data, NLGEOM=ON)

Greg

 

[Stringmaker]

I'm not sure what element type it is in Abaqus but there should be a solid element with a shell formulation that will make meshing much easier as you only need one element through the thickness. The Ansys version of this element supports just about any material model there is.

 

[brep]

feaplastic, did you ever resolve this issue? In particular, did you work out what you are doing wrong with the Abaqus shell model? As I previously stated, I can run this with 100% success...

 

[ActiveSolar05]

Have you tried bi-linear bricks??