where to do this…in prepomax?
When you create a new model, there are buttons to choose the model space. Just switch from 3D to 2D Plane Stress there (but the geometry has to lie completely on the XY plane):
Got it.. Thank You!!
Which measure of Alpha is finally requireed to be used in Prepomax? (Based on Initial (Linear) or actual length (Logarithmic)?
Derived strains (.0023) would have different meaning.
I remember Don Guido changed something in this regard ¿right?
PrePoMax uses the linear definition of ‘α’ based on initial length. So ε = α·ΔT. For small strains like 0.0023, linear and logarithmic are nearly the same. Not sure if Don Guido changed anything recently—would need to check.
There’s Guido’s explanation here: Thermal expansion with beam bending - #26 by dhondt - CalculiX (official versions are on www.calculix.de, the official GitHub repository is at https://github.com/Dhondtguido/CalculiX).
@FEAnalyst Earlier, you used a plane stress model for the analysis. However, when I tried the same approach for a composite analysis, the model showed errors related to S8 or S8R elements. Because of that, I switched to a 3D shell model for the composite laminate, which ran without errors.
But the results—especially the deflection and the stresses in individual laminate layers—didn’t match my hand calculations. I’m attaching the file for reference.
Thermal_composite.pmx (6.2 MB)
Could you please review it and suggest any corrections or improvements?
Yes, plane stress works for non-composite parts. For composite shell sections, you need 3D model.
Hard to say without seeing those hand calculations, but you should start from the simplest model and add complexity gradually. For example:
- make sure you get expected results with non-composite shell
- try with isotropic thermal expansion
- apply only simple mechanical load instead of thermal expansion (try different BCs if needed)
- simulate only one layer (make sure that its orientation is correct)
“deflection and the stresses in individual laminate layers—didn’t match my hand calculations”
Is your model based on some known benchmark?
You talk about 8 layers with different orientations, deflection involved, Poisson effect involved and orthotropic behaviour for the thermal expansion ..and you have compute it by hand ?¿?
¿Could I take a look at your results?
Please ignore my earlier message—it was sent in confusion. I now have a clear understanding of what I need.
Thank you for your suggestions!!
This could be a good reference.
1 Like