SPC Constraints ignored during composite optimization iteration steps
Dear all,
I am trying to set up a Composite optimization of the ply thicknesses of a single (curved) surface.
Part of the surface is a ring, modelled by a RBE2 or RBE3-Element which represents a pin which transfers a force into the part.
On the other edge of the surface are 5 holes, modelled by RBE2-Elements, which represent a screwed connection to another part. Those are my SPCs, where my loads are being transferred to. The part is symmetric, so that ultimately, there are 10 SPCs.
A Laminate made from multiple different orientation angles is applied to the surface, so the solver can minimize the thickness of plies not needed. A ply is modelled as orthotropic.
The applied tensile force is about 17000N, the compression force is about 4000N.
I applied the following optimization constraints:
- Displacement of force entry node of -0.2 to 0.2 (I have one tensile and one compression load case)
- The important one: Forces in SPCs between -2000N and -100N. (Only applicable to the tensile load step)
However, as the optimization starts, I would except the formation of paths from force entry Point to the SPCs, so that each SPC may carry the same load.
There are paths being formed, but unfortunately not enough. In most cases (which is dependent on the optimization constraints and ply thickness), only 4 of the desired paths form.
While Optistruct easily satisfies the displacement constraint, it does not take into account the Change in SPC Forces during the optimization.
Ultimately, 4 SPCs receive nearly all of the loads, while the remaining 6 do not have path leading to them and thus very low SPC Forces.
This was never an issue with isotropic metal optimization.
If neccessary, I could attach the .out-file, so that you may see what happens during the optimization, unless this is a known issue.
Answers
-
Hi,
Please share the .out file.
0 -
Hello,
I attached the OUT-File.
I have been talking to an Optistruct expert and she told me that my issue might have to do with the absense of CBUSH Elements when modelling the bolts. Could this be an issue? I assumed this was not neccessary if I am not making a frequency Analysis.
0 -
Okay I figured it out. Appeartently it has to do with a bug in the Version I am using (14.0.220). However, using CBUSHes bypassed this Problem and I am able to to the Simulation as intended.
0 -
Hi,
2017.2 is available to download. Please update to the same for new features and bug fixes.
0
Altair Employee