Single Layer of Element

[Lee.Conti]

Hi,

I have an assembly model, some parts are really thin. Instead of using Shell, some of us using solid but only single layer elements or the most two layers. I do not see any bad stress but is the stress reliable?

Also, when come to using tet10, does it still matter if you want to multiple layer of elements in thickness direction? Thank you!

 

[FEA way]

It depends on the type of load but generally it's not a good idea to leave parts with such a coarse mesh. This may cause many errors and spurious effects, some of which are hard to detect. Actually, too coarse mesh is one of the main reasons of errors and nonconvergence in numerical simulations. Trust me, I've seen many weird numerical issues that were solved when mesh was refined. Of course meshes made of second order elements have significantly higher accuracy and can be coarser but you still have to be very careful when it comes to mesh density.

 

[rb1957]

I'd consider it bad practice, as the solids would probably have very bad aspect ratios … unless you have bizillions of them (tiny t^3 bricks).

You'd also have "trouble" understanding any bending … you'd have to post-process, looking at the couple from nodal forces; and this is almost certainly a highly inaccurate method.

as for TET10 vs TET4 … don't use TET4s unless you Have to.

 

[Lee.Conti]

Thanks FEA way and rb1957.

The thin part was initially using hex8 but too difficult to model it with 2 elements across thickness and it would be even better to have 3 elements across the thickness... Tet10 was used instead of hex8.

 

[rb1957]

thin structures should, IMHO, use 2D elements.

another day in paradise, or is paradise one day closer ?

 

[SWComposites]

Hex8’s are very bad for shell bending. Either use 2d shell elements or 3d solid shells (such as available in Abaqus). And stop auto meshing CAD models of thin shell structures with solid elements unless you are using proper solid shell elements.

 

[Lee.Conti]

I do agree with both of you ,rb1957 & SWComposites.

I would say it was being lazy to do so in an assembly model to deal with the gap appears as a result of mid surface for fasteners and stack up contact locations.

 

[nlgyro]

I think you'll find the attached paper pertinent to your question. The summary mentioned in page 9 is interesting and depending on the solver you're using should help you in your modeling choices.

 

[rb1957]

"I would say it was being lazy to do so in an assembly model to deal with the gap appears as a result of mid surface for fasteners and stack up contact locations."

it's a sign (IMO unfortunate) of the times that we fret over t/2 gaps in models. Sure there are times when that is important, but not IMO for typical stress/load FEMs.

another day in paradise, or is paradise one day closer ?

 

[Lee.Conti]

Thank you nlgyro