Hello,
I’m trying to model a sandwich panel, core with skins each side, as solid core and shell skins but appear to have issues with tie constraints I cannot figure out. Hope that someone here could give me a hint what is going wrong?
Effectively the model is very similar as described here by @FEAnalyst and also mentioned in other post on this forum:
It all looks pretty straightforward, and I guess I overlooked something obvious, but somehow I cannot get the tie constraints working. It appears that the core and the skins are able to move independently.
It is a very basic model, both shells and the solid core are meshed with the same parameters, I played with the skin surface normals (pointing both towards core or pointing both away or both pointing in the same direction etc). The plate has a single pinned boundary condition on one side and a roller boundary condition on the opposite side (e.g. as found in mechanical testing) and a distributed pressure is applied on one skin. The tie constraints are created using “Find contact pairs” in the constraints section.
I do know that the tie constraints are the issue as I also tried to put fixed boundary condition between the edge of the skins and a side of the core. One can then clearly see that skins and core deflect independently of each other.
My set-up is here:
PrePoMax v2.2.10 def used.
Many thanks for any hint about what is going wrong and thank you all the same if you wouldn’t know either.
Thank you for this Jakub! Using the offset indeed fixed the issue. Prepomax is very user friendly and almost too intuitive which caught me out Therefore had to look up the offset keyword for Calculix to find out that this is a relative offset and not absolute:
Some software do have a little features which describes the input when hoovering over an input. Maybe some luxury addition for a future release to help “lazy” people like me out to get an answer quicker?!
From a workflow point of view, designing different offsets in CAD appears counterproductive. Using the offset within the Prepomax allows quicker iteration of different thicknesses also without the need of additional model imports and set-ups.
Unless I miss something, shouldn’t a tied shell (onto a solid) always be a 0.5 offset by default? I cannot imagine any situation where an overlap between parts could or should be required (therefore didn’t even consider the offset to begin with). Maybe this could be a feature request then: tied shells are always meshed at 0.5 offset?
Anyhow, thank you again for your comment, saved me some sleepless nights!
Yes understand the logic behind it (now). Just an observation to be consistent, should the offset in Thicken Shell Mesh then not be defined in the same way rather than in absolute distances?
Therefore had to look up the offset keyword for Calculix to find out that this is a relative offset and not absolute: