Thermal Stress in Shells using Laminate Material

kuhl

Hi all,
picking up a side note from my previous post:
https://mecway.com/forum/discussion/1765/element-orientation-in-solids#latest

Is there a way to consider thermal expansion when using shells with laminate material properties either with internal solver or CXX? The input does not seem to allow it.



When using CXX, could thermal expansion in thickness direction be considered as CXX expands shells to solids?

Thanks!

kuhl

I did some testing on a 100 mm cube:
left: shell model, right: solid model.
Element orientation identical to global coordinates.DeltaT=-100K


Shell material, one layer


Solid material


So far, no thermal expansion defined for shells, results as expeced:


If I introduce custom model definition for thermal expansion, so that the shell model should behave like the solid:


Result for the shell model is warped:


I added node rotation constraint to stop the warping although I thought this shouldnt be necessary:


Result: Displacement of shell and solid very similar:


changing the orientation of the layer definition:


works as expected, so thermal expansion in custom model definition does not seem to overwrite the thermal expansion defined in the orthotropic solid model.



so far so good. BUT...

if I add a second layer to the shells (orientation back to 0°)


I get this:


When using four layers, this happens:


looking at the stress, there is something happening between the top layer and the other layers of the shell:


releasing the rotation DOFS again to allow deformation and adding more layers, then looking at the stress again shows that the top layer is in tension (x) while all others are in compression (the solid is free of stress).


So it seems that the thermal expansion defined in the custom model definition is only acting on the top layer. Or maybe the reference temperatur is not applied to all layers in the same way...

Does anyone have an idea?

Victor

This looks like a bug Disla found https://github.com/Dhondtguido/CalculiX/issues/129 for heat flow which has apparently been fixed in CCX 2.23.

kuhl

Thanks, Victor!