Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Changing Constraints Every Iteration During Size Optimization

Hi,

I am doing a size optimization of a sandwich composite for thickness. The thing is, every time you change the thickness the stress constraints change with the design variables. Is there any way to address this issue in the software? For example, my stress constraint in a ply might be 423 MPa, but that stress constraint may drop if the thickness is changed to 300 MPa, so the stress constraint cannot be fixed.

Edit2: Also, if possible, how can you adjust stress constraints based on element size/area?

Thanks in advance!

Find more posts tagged with

English

OptiStruct

Accepted answers

All comments

Rahul Rajan_21763

Could you pleas make sure to keep elements that are part of the optimization design-space must have a separate property card from the non-design elements.

Crashphys

Hi Rahul,

Can you please clarify? Each structural element I have has its own property. I have 7 properties I am optimizing for, each with their own stress constraints. The thing is, for 5 of them my stress constraint is equal to some multiple of their thickness, so their stress constraint changes with the design variables. Is it possible to relate a design variable to the stress cosntraints?

Robert Lawson

Crashphys -

You can't accomplish exactly that, but here's a work-around. Create a DRESP1 with your stress response. Then create a DRESP2 with DEQATN. Pass the stress response and the property thickness (assuming this is a DESVAR) to the DEQATN. I'm not sure if you intend a linear relationship or something else. But your DRESP2/DEQATN could be written such that your stress responses is adjusted for thickness (or some function of thickness). Then constrain that stress with DCONSTR.

- Robert

Crashphys

Thank you very much for the suggestion. I will try it an report back for future users.

Crashphys

In case anyone is ever wondering, I finally got around to trying this and Robert's answer works.

Thanks Robert!