shell coat on solid elements

[james80]

All
I wanted to know why it is recommended to put a shell coat on solid elements?
One possible reason could be that there is a degree of freedom mismatch if the FEA model includes shell-solid connections.
Is there any other reason
I appreciate your help
thanks
regards
James

 

[johnhors]

James,

Who made the recommendation and in what context?
As you have already indicated there is a compatibility issue when mixing elements of different types which use different shape and displacement functions. It is not good practice to do this and can give rise to errors in result interpretations. I know many analysts do skin solids with shell or membrane elements, but I have yet to see an argument for this practice that holds up under scrutiny.

 

[james80]

John
Thanks for your reply.
By shell coat I meant membrane elements.
I apologise for the confusion.

james

 

[jetmaker]

I have used this practice for reworking solids to remove hotspots by adding effective mass. Where this application was used is when design and modelling iterations were costly, so you would skin the solid to increase "effective thickness". Once a few analysis iterations were complet, the skin thickness increase was passed on to the designer to update the solid geometry. This would then be used in another non-skinned FEA to confirm that the hotspots had been addressed.

jetmaker.

 

[johnhors]

Jetmaker,

Yes, I see your point using it as a quick, crude design iteration, but as you say you don't trust your final analysis being skinned.

I'd still like to know where James has seen this as a "recommendation" !

 

[GregLocock]

I know people who put a zero thickness skin on a solid element so they can find the extreme stress values.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[brep]

It's a common, accepted practice where the surface stresses are of interest. The model is, however, might be meshed with solids tets,which do not have gauss points on the outer surface (results will be extrapolated). By putting a layer of membranes on the surface of the solids, you get gauss points precisely where you need them for data output, i.e. on the surface.

 

[corus]

I agree with others regarding the use of shell elements for surface stresses, but I'm not so sure about a zero thickness. Surely that would give an error as the stiffness matrix would have a zero diagonal? Equally I heard that having two elements adjacent to each other where there was a significant difference in the stiffness would also lead to greater inaccuracies and hence the need for a gradual change in mesh density when modelling. I'm not sure if having shell elements with a small, but non-zero, thickness adjacent to solid elements would therefore be of benefit. Any references showing the benefit would be useful.

corus

 

[GregLocock]

Sorry, I meant notional rather than zero.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[james80]

All
thanks for the responses.
I appreicate your help
james80